diff --git a/alembic/versions/2026_03_31_1915-5516c5bf7762_add_run_history_columns_to_task_runs.py b/alembic/versions/2026_03_31_1915-5516c5bf7762_add_run_history_columns_to_task_runs.py new file mode 100644 index 000000000..b12142900 --- /dev/null +++ b/alembic/versions/2026_03_31_1915-5516c5bf7762_add_run_history_columns_to_task_runs.py @@ -0,0 +1,81 @@ +"""add run history columns to task_runs + +Revision ID: 5516c5bf7762 +Revises: d77ed2605df0 +Create Date: 2026-03-31T19:15:14.171986+00:00 + +""" + +from typing import Sequence, Union + +import sqlalchemy as sa + +from alembic import op + +# revision identifiers, used by Alembic. +revision: str = "5516c5bf7762" +down_revision: Union[str, None] = "d77ed2605df0" +branch_labels: Union[str, Sequence[str], None] = None +depends_on: Union[str, Sequence[str], None] = None + + +def upgrade() -> None: + op.add_column("task_runs", sa.Column("status", sa.String(), nullable=True)) + op.add_column("task_runs", sa.Column("started_at", sa.DateTime(), nullable=True)) + op.add_column("task_runs", sa.Column("finished_at", sa.DateTime(), nullable=True)) + op.add_column("task_runs", sa.Column("workflow_permanent_id", sa.String(), nullable=True)) + op.add_column("task_runs", sa.Column("script_run", sa.JSON(), nullable=True)) + op.add_column("task_runs", sa.Column("parent_workflow_run_id", sa.String(), nullable=True)) + op.add_column("task_runs", sa.Column("debug_session_id", sa.String(), nullable=True)) + op.add_column("task_runs", sa.Column("searchable_text", sa.Text(), nullable=True)) + # pg_trgm is available by default on RDS and Cloud SQL. + op.execute("CREATE EXTENSION IF NOT EXISTS pg_trgm") + # All indexes are created CONCURRENTLY to avoid SHARE locks that block writes + # on the actively-written task_runs table. + with op.get_context().autocommit_block(): + op.execute("SET statement_timeout = '3h';") + try: + op.execute( + "CREATE INDEX CONCURRENTLY IF NOT EXISTS ix_task_runs_org_status_created " + "ON task_runs USING btree (organization_id, status, created_at DESC)" + ) + op.execute( + "CREATE INDEX CONCURRENTLY IF NOT EXISTS ix_task_runs_searchable_text_gin " + "ON task_runs USING gin (searchable_text gin_trgm_ops)" + ) + op.execute(""" + CREATE INDEX CONCURRENTLY IF NOT EXISTS ix_task_runs_org_toplevel_created + ON task_runs (organization_id, created_at DESC) + WHERE parent_workflow_run_id IS NULL + AND debug_session_id IS NULL + AND status IS NOT NULL; + """) + # Partial index covering non-terminal task_runs rows. + # Used by the task_runs_sync_activity cron to efficiently find rows + # that still need syncing. The index shrinks as runs complete. + op.execute(""" + CREATE INDEX CONCURRENTLY IF NOT EXISTS ix_task_runs_nonterminal + ON task_runs (run_id, task_run_type) + WHERE status IS NULL + OR status NOT IN ('completed', 'failed', 'terminated', 'canceled', 'timed_out') + """) + finally: + try: + op.execute("RESET statement_timeout;") + except Exception: + pass # Don't mask the original index creation error + + +def downgrade() -> None: + op.execute("DROP INDEX IF EXISTS ix_task_runs_nonterminal") + op.execute("DROP INDEX IF EXISTS ix_task_runs_org_toplevel_created") + op.execute("DROP INDEX IF EXISTS ix_task_runs_searchable_text_gin") + op.execute("DROP INDEX IF EXISTS ix_task_runs_org_status_created") + op.drop_column("task_runs", "searchable_text") + op.drop_column("task_runs", "debug_session_id") + op.drop_column("task_runs", "parent_workflow_run_id") + op.drop_column("task_runs", "script_run") + op.drop_column("task_runs", "workflow_permanent_id") + op.drop_column("task_runs", "finished_at") + op.drop_column("task_runs", "started_at") + op.drop_column("task_runs", "status") diff --git a/pyproject.toml b/pyproject.toml index e9ba25c1f..b1984da93 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -167,6 +167,9 @@ include = [ "skyvern-frontend/tsconfig.node.json" = "skyvern-frontend/tsconfig.node.json" [tool.uv] +# Supply chain quarantine: block packages published less than 7 days ago. +# Override for urgent cases: uv add --exclude-newer "" +exclude-newer = "7 days" constraint-dependencies = [ "authlib>=1.6.9", "flask>=3.1.3", diff --git a/run_alembic_check.sh b/run_alembic_check.sh index 279264bc2..3d20eda44 100755 --- a/run_alembic_check.sh +++ b/run_alembic_check.sh @@ -1,4 +1,10 @@ #!/bin/sh +# Alembic migrations are PostgreSQL-specific. Default to PostgreSQL when +# DATABASE_STRING is not set (e.g. OSS CI), so we don't pick up the new +# SQLite default from Settings._default_database_string(). +: "${DATABASE_STRING:=postgresql+asyncpg://skyvern@localhost/skyvern}" +export DATABASE_STRING + # first apply migrations export PATH="${PATH}:.venv/bin" alembic upgrade head diff --git a/skyvern-frontend/src/api/types.ts b/skyvern-frontend/src/api/types.ts index b8d1a0626..bab46af3d 100644 --- a/skyvern-frontend/src/api/types.ts +++ b/skyvern-frontend/src/api/types.ts @@ -441,6 +441,31 @@ export type WorkflowRunApiResponse = { workflow_title: string | null; }; +export const TaskRunType = { + TaskV1: "task_v1", + TaskV2: "task_v2", + WorkflowRun: "workflow_run", + OpenaiCua: "openai_cua", + AnthropicCua: "anthropic_cua", + UiTars: "ui_tars", +} as const; + +export type TaskRunType = (typeof TaskRunType)[keyof typeof TaskRunType]; + +export type TaskRunListItem = { + task_run_id: string; + task_run_type: TaskRunType; + run_id: string; + title: string | null; + status: string; + started_at: string | null; + finished_at: string | null; + created_at: string; + workflow_permanent_id: string | null; + script_run: boolean; + searchable_text: string | null; +}; + export type WorkflowRunStatusApiResponse = { workflow_id: string; workflow_run_id: string; diff --git a/skyvern-frontend/src/hooks/useRunsQuery.ts b/skyvern-frontend/src/hooks/useRunsQuery.ts index 2ea9cb500..66f49866b 100644 --- a/skyvern-frontend/src/hooks/useRunsQuery.ts +++ b/skyvern-frontend/src/hooks/useRunsQuery.ts @@ -1,9 +1,9 @@ import { getClient } from "@/api/AxiosClient"; import { useCredentialGetter } from "@/hooks/useCredentialGetter"; import { useQuery } from "@tanstack/react-query"; -import { Status, Task, TriggerType, WorkflowRunApiResponse } from "@/api/types"; +import { Status, TaskRunListItem } from "@/api/types"; -type QueryReturnType = Array; +type QueryReturnType = Array; type UseQueryOptions = Omit< Parameters>[0], "queryKey" | "queryFn" @@ -13,7 +13,6 @@ type Props = { page?: number; pageSize?: number; statusFilters?: Array; - triggerTypeFilters?: Array; search?: string; } & UseQueryOptions; @@ -21,20 +20,14 @@ function useRunsQuery({ page = 1, pageSize = 10, statusFilters, - triggerTypeFilters, search, + ...queryOptions }: Props) { const credentialGetter = useCredentialGetter(); - return useQuery>({ - queryKey: [ - "runs", - { statusFilters, triggerTypeFilters }, - page, - pageSize, - search, - ], + return useQuery>({ + queryKey: ["runs", { statusFilters }, page, pageSize, search], queryFn: async () => { - const client = await getClient(credentialGetter); + const client = await getClient(credentialGetter, "sans-api-v1"); const params = new URLSearchParams(); params.append("page", String(page)); params.append("page_size", String(pageSize)); @@ -43,16 +36,12 @@ function useRunsQuery({ params.append("status", status); }); } - if (triggerTypeFilters) { - triggerTypeFilters.forEach((triggerType) => { - params.append("trigger_type", triggerType); - }); - } if (search) { params.append("search_key", search); } return client.get("/runs", { params }).then((res) => res.data); }, + ...queryOptions, }); } diff --git a/skyvern-frontend/src/routes/history/RunHistory.tsx b/skyvern-frontend/src/routes/history/RunHistory.tsx index 245adc039..43217c24e 100644 --- a/skyvern-frontend/src/routes/history/RunHistory.tsx +++ b/skyvern-frontend/src/routes/history/RunHistory.tsx @@ -3,15 +3,12 @@ import { LightningBoltIcon, MixerHorizontalIcon } from "@radix-ui/react-icons"; import { Tip } from "@/components/Tip"; import { Status, - Task, - TriggerType, - WorkflowRunApiResponse, + TaskRunListItem, + TaskRunType, WorkflowRunStatusApiResponse, } from "@/api/types"; import { StatusBadge } from "@/components/StatusBadge"; import { StatusFilterDropdown } from "@/components/StatusFilterDropdown"; -import { TriggerTypeBadge } from "@/components/TriggerTypeBadge"; -import { TriggerTypeFilterDropdown } from "@/components/TriggerTypeFilterDropdown"; import { Pagination, PaginationContent, @@ -41,7 +38,6 @@ import React, { useEffect, useState } from "react"; import { useNavigate, useSearchParams } from "react-router-dom"; import { getClient } from "@/api/AxiosClient"; import { useCredentialGetter } from "@/hooks/useCredentialGetter"; -import * as env from "@/util/env"; import { useDebounce } from "use-debounce"; import { Button } from "@/components/ui/button"; import { @@ -57,21 +53,33 @@ import { useParameterExpansion } from "@/routes/workflows/hooks/useParameterExpa import { ParameterDisplayInline } from "@/routes/workflows/components/ParameterDisplayInline"; import { HighlightText } from "@/routes/workflows/components/HighlightText"; -function isTask(run: Task | WorkflowRunApiResponse): run is Task { - return "task_id" in run; +const statusValues = new Set(Object.values(Status)); +function isKnownStatus(value: string): value is Status { + return statusValues.has(value); +} + +function getRunNavigationPath(run: TaskRunListItem): string { + switch (run.task_run_type) { + case TaskRunType.WorkflowRun: + case TaskRunType.TaskV2: + return `/runs/${run.run_id}`; + case TaskRunType.TaskV1: + case TaskRunType.OpenaiCua: + case TaskRunType.AnthropicCua: + case TaskRunType.UiTars: + return `/tasks/${run.run_id}/actions`; + default: + return `/runs/${run.run_id}`; + } } function RunHistory() { - const credentialGetter = useCredentialGetter(); const [searchParams, setSearchParams] = useSearchParams(); const page = searchParams.get("page") ? Number(searchParams.get("page")) : 1; const itemsPerPage = searchParams.get("page_size") ? Number(searchParams.get("page_size")) : 10; const [statusFilters, setStatusFilters] = useState>([]); - const [triggerTypeFilters, setTriggerTypeFilters] = useState< - Array - >([]); const [search, setSearch] = useState(""); const [debouncedSearch] = useDebounce(search, 500); @@ -79,43 +87,17 @@ function RunHistory() { page, pageSize: itemsPerPage, statusFilters, - triggerTypeFilters, search: debouncedSearch, }); const navigate = useNavigate(); - const { data: nextPageRuns } = useQuery>( - { - queryKey: [ - "runs", - { statusFilters, triggerTypeFilters }, - page + 1, - itemsPerPage, - debouncedSearch, - ], - queryFn: async () => { - const client = await getClient(credentialGetter); - const params = new URLSearchParams(); - params.append("page", String(page + 1)); - params.append("page_size", String(itemsPerPage)); - if (statusFilters) { - statusFilters.forEach((status) => { - params.append("status", status); - }); - } - if (triggerTypeFilters) { - triggerTypeFilters.forEach((triggerType) => { - params.append("trigger_type", triggerType); - }); - } - if (debouncedSearch) { - params.append("search_key", debouncedSearch); - } - return client.get("/runs", { params }).then((res) => res.data); - }, - enabled: runs && runs.length === itemsPerPage, - }, - ); + const { data: nextPageRuns } = useRunsQuery({ + page: page + 1, + pageSize: itemsPerPage, + statusFilters, + search: debouncedSearch, + enabled: runs?.length === itemsPerPage, + }); const isNextDisabled = isFetching || !nextPageRuns || nextPageRuns.length === 0; @@ -136,8 +118,8 @@ function RunHistory() { const workflowRunIds = runs - ?.filter((run): run is WorkflowRunApiResponse => !isTask(run)) - .map((run) => run.workflow_run_id) + ?.filter((run) => run.task_run_type === TaskRunType.WorkflowRun) + .map((run) => run.run_id) .filter((id): id is string => Boolean(id)) ?? []; setAutoExpandedRows(workflowRunIds); @@ -170,12 +152,137 @@ function RunHistory() { if (isNextDisabled) return; setParamPatch({ page: String(page + 1) }); } + + const displayTableBody = () => { + // Show loading skeleton + if (isFetching) { + return Array.from({ length: 10 }).map((_, index) => ( + + + + + + )); + } + + // No runs found + if (runs?.length === 0) { + return ( + + +
No runs found
+ + + ); + } + + return runs?.map((run) => { + const executionTime = formatExecutionTime( + run.started_at ?? run.created_at, + run.finished_at, + ); + const isWorkflowRun = run.task_run_type === TaskRunType.WorkflowRun; + const isExpanded = isWorkflowRun && expandedRows.has(run.run_id); + const navPath = getRunNavigationPath(run); + + const titleContent = run.script_run ? ( +
+ + + + {run.title ?? ""} +
+ ) : ( + run.title ?? "" + ); + + return ( + + { + handleNavigate(event, navPath); + }} + > + + + + + {titleContent} + + + {isKnownStatus(run.status) ? ( + + ) : ( + {run.status} + )} + + + {basicLocalTimeFormat(run.created_at)} + + + {executionTime ?? "-"} + + + {isWorkflowRun ? ( +
+ + + + + + + {isExpanded ? "Hide Parameters" : "Show Parameters"} + + + +
+ ) : null} + + + + {isExpanded && run.workflow_permanent_id && ( + + + + + + )} + + ); + }); + }; + return (

Run History

-
+
{ @@ -187,19 +294,10 @@ function RunHistory() { placeholder="Search by run ID or parameter..." className="w-48 lg:w-72" /> - { - setTriggerTypeFilters(values); - const params = new URLSearchParams(searchParams); - params.set("page", "1"); - setSearchParams(params, { replace: true }); - }} - /> { - setStatusFilters(values); + onChange={(filters) => { + setStatusFilters(filters); const params = new URLSearchParams(searchParams); params.set("page", "1"); setSearchParams(params, { replace: true }); @@ -222,184 +320,7 @@ function RunHistory() { - - {isFetching ? ( - Array.from({ length: 10 }).map((_, index) => ( - - - - - - )) - ) : runs?.length === 0 ? ( - - -
No runs found
- - - ) : ( - runs?.map((run) => { - if (isTask(run)) { - const taskExecutionTime = formatExecutionTime( - run.started_at ?? run.created_at, - run.finished_at, - ); - return ( - { - handleNavigate(event, `/tasks/${run.task_id}/actions`); - }} - > - - {run.task_id} - - - {run.url} - - - - - - {basicLocalTimeFormat(run.created_at)} - - - {taskExecutionTime ?? "-"} - - {/* Align with workflow row's expand button column. */} - - - ); - } - - const workflowTitle = ( -
- {run.workflow_title ?? ""} - {run.script_run === true && ( - - - - )} - -
- ); - - const isExpanded = expandedRows.has(run.workflow_run_id); - const workflowExecutionTime = formatExecutionTime( - run.started_at ?? run.created_at, - run.finished_at, - ); - - return ( - - { - handleNavigate( - event, - env.useNewRunsUrl - ? `/runs/${run.workflow_run_id}` - : `/workflows/${run.workflow_permanent_id}/${run.workflow_run_id}/overview`, - ); - }} - > - - - - - {workflowTitle} - - - - - -
- {basicLocalTimeFormat(run.created_at)} - {run.trigger_type === TriggerType.Scheduled && - run.schedule_name && ( - - {run.schedule_name} - {run.schedule_cron - ? ` (${run.schedule_cron})` - : ""} - - )} -
- - - {workflowExecutionTime ?? "-"} - - -
- - - - - - - {isExpanded - ? "Hide Parameters" - : "Show Parameters"} - - - -
- - - - {isExpanded && ( - - - - - - )} - - ); - }) - )} - + {displayTableBody()}
@@ -466,7 +387,7 @@ function WorkflowRunParametersInline({ workflowRunId, searchQuery, keywordMatchesParameter, -}: WorkflowRunParametersInlineProps) { +}: Readonly) { const { data: globalWorkflows } = useGlobalWorkflowsQuery(); const credentialGetter = useCredentialGetter(); diff --git a/skyvern/config.py b/skyvern/config.py index 25ceb69d8..abea0005d 100644 --- a/skyvern/config.py +++ b/skyvern/config.py @@ -1,13 +1,43 @@ import logging import platform +from pathlib import Path from typing import Any +from pydantic import Field from pydantic_settings import BaseSettings, SettingsConfigDict from skyvern import constants from skyvern.constants import REPO_ROOT_DIR, SKYVERN_DIR from skyvern.utils.env_paths import resolve_backend_env_path + +def _default_database_string() -> str: + """Return the default DATABASE_STRING. + + Uses a SQLite file at ~/.skyvern/data.db so that ``skyvern run server`` + works out of the box without Docker or Postgres. Users who set + DATABASE_STRING in .env or the environment get Postgres automatically + (pydantic-settings reads env before the default_factory runs). + + This is a pure string computation — no filesystem side effects. + The parent directory is created by _ensure_sqlite_dir() at engine + build time (agent_db.py) or server bootstrap time (api_app.py). + """ + db_path = Path.home() / ".skyvern" / "data.db" + return f"sqlite+aiosqlite:///{db_path}" + + +def _ensure_sqlite_dir(database_string: str) -> None: + """Create the parent directory for a file-backed SQLite database URL. + + No-op for in-memory SQLite (`:memory:`) or non-SQLite URLs. + """ + if not database_string.startswith("sqlite") or ":memory:" in database_string: + return + db_file = database_string.split("///", 1)[-1] + Path(db_file).expanduser().resolve().parent.mkdir(parents=True, exist_ok=True) + + # NOTE: _DEFAULT_ENV_FILES resolves .env paths at import time and assumes # the process has changed dir to the desired project root by this time. # Even if we were to resolve paths at instantiation time, the global `settings` @@ -64,11 +94,7 @@ class Settings(BaseSettings): LONG_RUNNING_TASK_WARNING_RATIO: float = 0.95 MAX_RETRIES_PER_STEP: int = 5 DEBUG_MODE: bool = False - DATABASE_STRING: str = ( - "postgresql+asyncpg://skyvern@localhost/skyvern" - if platform.system() == "Windows" - else "postgresql+psycopg://skyvern@localhost/skyvern" - ) + DATABASE_STRING: str = Field(default_factory=_default_database_string) DATABASE_REPLICA_STRING: str | None = None DATABASE_STATEMENT_TIMEOUT_MS: int = 60000 DISABLE_CONNECTION_POOL: bool = False @@ -630,6 +656,9 @@ class Settings(BaseSettings): ) object.__setattr__(self, "DATABASE_STRING", updated_string) + def is_sqlite(self) -> bool: + return self.DATABASE_STRING.startswith("sqlite") + def is_cloud_environment(self) -> bool: """ :return: True if env is not local, else False diff --git a/skyvern/exceptions.py b/skyvern/exceptions.py index e7fa8639f..9a728dc12 100644 --- a/skyvern/exceptions.py +++ b/skyvern/exceptions.py @@ -895,8 +895,14 @@ class FailedToGetTOTPVerificationCode(SkyvernException): class SkyvernContextWindowExceededError(SkyvernException): - def __init__(self) -> None: - message = "Context window exceeded. Please contact support@skyvern.com for help." + def __init__(self, model: str | None = None, prompt_name: str | None = None) -> None: + details = [] + if model: + details.append(f"model: {model}") + if prompt_name: + details.append(f"prompt: {prompt_name}") + detail_str = f" ({', '.join(details)})" if details else "" + message = f"LLM context window exceeded{detail_str}. The page may have too much content for the AI model to process. Please try again or contact support@skyvern.com for help." super().__init__(message) diff --git a/skyvern/forge/api_app.py b/skyvern/forge/api_app.py index 696ba03cc..c09de626d 100644 --- a/skyvern/forge/api_app.py +++ b/skyvern/forge/api_app.py @@ -13,7 +13,7 @@ from starlette.requests import HTTPConnection, Request from starlette_context.middleware import RawContextMiddleware from starlette_context.plugins.base import Plugin -from skyvern.config import settings +from skyvern.config import _ensure_sqlite_dir, settings from skyvern.exceptions import SkyvernHTTPException from skyvern.forge import app as forge_app from skyvern.forge.forge_app_initializer import start_forge_app @@ -21,8 +21,15 @@ from skyvern.forge.request_logging import log_raw_request_middleware from skyvern.forge.sdk.core import skyvern_context from skyvern.forge.sdk.core.skyvern_context import SkyvernContext from skyvern.forge.sdk.db.exceptions import NotFoundError +from skyvern.forge.sdk.db.models import Base from skyvern.forge.sdk.routes import internal_auth from skyvern.forge.sdk.routes.routers import base_router, legacy_base_router, legacy_v2_router +from skyvern.forge.sdk.services.local_org_auth_token_service import ( + ensure_local_api_key, + ensure_local_org, + fingerprint_token, + regenerate_local_api_key, +) from skyvern.services.cleanup_service import start_cleanup_scheduler, stop_cleanup_scheduler LOG = structlog.get_logger() @@ -74,11 +81,65 @@ def custom_openapi(app: FastAPI) -> dict: return app.openapi_schema +async def _bootstrap_sqlite() -> None: + """Auto-bootstrap SQLite on first server start. + + Creates tables, a local org, and an API key so that + ``skyvern run server`` works out of the box with zero configuration. + Idempotent: skips if the org already exists. + """ + _ensure_sqlite_dir(settings.DATABASE_STRING) + + db = forge_app.DATABASE + async with db.engine.begin() as conn: + await conn.run_sync(Base.metadata.create_all) + + # Preserve an existing API key if it's a real value (not the skeleton default). + # settings.SKYVERN_API_KEY already incorporates env vars and .env via pydantic-settings. + existing_key = settings.SKYVERN_API_KEY if settings.SKYVERN_API_KEY != "PLACEHOLDER" else None + + if existing_key: + preserved = await ensure_local_api_key(existing_key) + if preserved is not None: + api_key, org_id = preserved + LOG.info( + "Existing SKYVERN_API_KEY detected — preserving env value and syncing it into the local SQLite DB.", + organization_id=org_id, + api_key_fingerprint=fingerprint_token(api_key), + ) + return + + LOG.warning( + "Existing SKYVERN_API_KEY could not be preserved for local SQLite bootstrap; generating a new local key.", + ) + + organization = await ensure_local_org() + existing_token = await db.get_valid_org_auth_token(organization.organization_id, "api") + if existing_token is not None: + LOG.info("SQLite database already bootstrapped", organization_id=organization.organization_id) + return + + api_key, org_id, backend_env, frontend_env = await regenerate_local_api_key() + LOG.info( + "SQLite bootstrap complete — local org and API key created", + organization_id=org_id, + api_key_fingerprint=fingerprint_token(api_key), + env_file_written=backend_env, + ) + + @asynccontextmanager async def lifespan(fastapi_app: FastAPI) -> AsyncGenerator[None, Any]: """Lifespan context manager for FastAPI app startup and shutdown.""" LOG.info("Server started") + + # Auto-bootstrap SQLite database on first server start. + # Re-raise on failure — a server with no tables/org/API key is + # useless and would produce confusing 401s on every request. + if settings.is_sqlite(): + await _bootstrap_sqlite() + if forge_app.api_app_startup_event: LOG.info("Calling api app startup event") try: diff --git a/skyvern/forge/sdk/api/llm/api_handler_factory.py b/skyvern/forge/sdk/api/llm/api_handler_factory.py index 02a55309b..91b14adcb 100644 --- a/skyvern/forge/sdk/api/llm/api_handler_factory.py +++ b/skyvern/forge/sdk/api/llm/api_handler_factory.py @@ -766,7 +766,7 @@ class LLMAPIHandlerFactory: prompt_name=prompt_name, duration_seconds=duration_seconds, ) - raise SkyvernContextWindowExceededError() from e + raise SkyvernContextWindowExceededError(model=main_model_group, prompt_name=prompt_name) from e except ValueError as e: duration_seconds = time.time() - start_time LOG.exception( @@ -1217,7 +1217,7 @@ class LLMAPIHandlerFactory: prompt_name=prompt_name, duration_seconds=duration_seconds, ) - raise SkyvernContextWindowExceededError() from e + raise SkyvernContextWindowExceededError(model=model_name, prompt_name=prompt_name) from e except CancelledError: # Speculative steps are intentionally cancelled when goal verification completes first, # so we log at debug level. Non-speculative cancellations are unexpected errors. @@ -1664,7 +1664,7 @@ class LLMCaller: llm_key=self.llm_key, model=self.llm_config.model_name, ) - raise SkyvernContextWindowExceededError() from e + raise SkyvernContextWindowExceededError(model=self.llm_config.model_name) from e except CancelledError: # Speculative steps are intentionally cancelled when goal verification returns completed, # so we log at debug level. Non-speculative cancellations are unexpected errors. diff --git a/skyvern/forge/sdk/api/llm/exceptions.py b/skyvern/forge/sdk/api/llm/exceptions.py index 70f602547..896189170 100644 --- a/skyvern/forge/sdk/api/llm/exceptions.py +++ b/skyvern/forge/sdk/api/llm/exceptions.py @@ -51,7 +51,8 @@ class LLMProviderError(BaseLLMError): class LLMProviderErrorRetryableTask(LLMProviderError): def __init__(self, llm_key: str) -> None: - super().__init__(f"Retryable error while using LLMProvider {llm_key}") + # Call BaseLLMError directly to avoid double-formatting the message through LLMProviderError.__init__ + BaseLLMError.__init__(self, f"Retryable error while using LLMProvider {llm_key}") class NoProviderEnabledError(BaseLLMError): diff --git a/skyvern/forge/sdk/db/agent_db.py b/skyvern/forge/sdk/db/agent_db.py index c8f921583..75a5ef980 100644 --- a/skyvern/forge/sdk/db/agent_db.py +++ b/skyvern/forge/sdk/db/agent_db.py @@ -54,6 +54,9 @@ def _build_engine(database_string: str) -> AsyncEngine: Always enables foreign key enforcement via PRAGMA. """ if database_string.startswith("sqlite"): + from skyvern.config import _ensure_sqlite_dir + + _ensure_sqlite_dir(database_string) is_memory = ":memory:" in database_string engine_kwargs: dict[str, Any] = { "json_serializer": _custom_json_serializer, @@ -327,6 +330,9 @@ class AgentDB(BaseAlchemyDB): async def get_all_runs(self, *args: Any, **kwargs: Any) -> Any: return await self.workflow_runs.get_all_runs(*args, **kwargs) + async def get_all_runs_v2(self, *args: Any, **kwargs: Any) -> Any: + return await self.workflow_runs.get_all_runs_v2(*args, **kwargs) + async def get_workflow_run(self, *args: Any, **kwargs: Any) -> Any: return await self.workflow_runs.get_workflow_run(*args, **kwargs) @@ -444,22 +450,25 @@ class AgentDB(BaseAlchemyDB): return await self.workflow_params.retrieve_action_plan(*args, **kwargs) async def create_task_run(self, *args: Any, **kwargs: Any) -> Any: - return await self.workflow_params.create_task_run(*args, **kwargs) + return await self.tasks.create_task_run(*args, **kwargs) async def update_task_run(self, *args: Any, **kwargs: Any) -> Any: - return await self.workflow_params.update_task_run(*args, **kwargs) + return await self.tasks.update_task_run(*args, **kwargs) + + async def sync_task_run_status(self, *args: Any, **kwargs: Any) -> Any: + return await self.tasks.sync_task_run_status(*args, **kwargs) async def update_job_run_compute_cost(self, *args: Any, **kwargs: Any) -> Any: - return await self.workflow_params.update_job_run_compute_cost(*args, **kwargs) + return await self.tasks.update_job_run_compute_cost(*args, **kwargs) async def cache_task_run(self, *args: Any, **kwargs: Any) -> Any: - return await self.workflow_params.cache_task_run(*args, **kwargs) + return await self.tasks.cache_task_run(*args, **kwargs) async def get_cached_task_run(self, *args: Any, **kwargs: Any) -> Any: - return await self.workflow_params.get_cached_task_run(*args, **kwargs) + return await self.tasks.get_cached_task_run(*args, **kwargs) async def get_run(self, *args: Any, **kwargs: Any) -> Any: - return await self.workflow_params.get_run(*args, **kwargs) + return await self.tasks.get_run(*args, **kwargs) # -- Artifact delegates -- diff --git a/skyvern/forge/sdk/db/agent_db_test.py b/skyvern/forge/sdk/db/agent_db_test.py index 94c551bf1..539e7099f 100644 --- a/skyvern/forge/sdk/db/agent_db_test.py +++ b/skyvern/forge/sdk/db/agent_db_test.py @@ -43,6 +43,22 @@ async def test_create_organization(agent_db: AgentDB) -> None: assert retrieved_by_domain.domain == domain +@pytest.mark.asyncio +async def test_create_organization_with_explicit_id(agent_db: AgentDB) -> None: + organization = await agent_db.create_organization( + organization_id="o_test_org", + organization_name="Explicit Id Organization", + domain="explicit.test", + ) + + assert organization.organization_id == "o_test_org" + + retrieved_org = await agent_db.get_organization("o_test_org") + assert retrieved_org is not None + assert retrieved_org.organization_name == "Explicit Id Organization" + assert retrieved_org.domain == "explicit.test" + + @pytest.mark.asyncio async def test_get_organization_not_found(agent_db: AgentDB) -> None: retrieved_org = await agent_db.get_organization("non_existent_id") diff --git a/skyvern/forge/sdk/db/mixins/organizations.py b/skyvern/forge/sdk/db/mixins/organizations.py index cb2b47104..dc95be82f 100644 --- a/skyvern/forge/sdk/db/mixins/organizations.py +++ b/skyvern/forge/sdk/db/mixins/organizations.py @@ -132,9 +132,11 @@ class OrganizationsMixin: max_steps_per_run: int | None = None, max_retries_per_step: int | None = None, domain: str | None = None, + organization_id: str | None = None, ) -> Organization: async with self.Session() as session: org = OrganizationModel( + organization_id=organization_id, organization_name=organization_name, webhook_callback_url=webhook_callback_url, max_steps_per_run=max_steps_per_run, diff --git a/skyvern/forge/sdk/db/mixins/tasks.py b/skyvern/forge/sdk/db/mixins/tasks.py index 01fef36b7..7f42ee438 100644 --- a/skyvern/forge/sdk/db/mixins/tasks.py +++ b/skyvern/forge/sdk/db/mixins/tasks.py @@ -14,13 +14,15 @@ from skyvern.forge.sdk.db.models import ( ActionModel, StepModel, TaskModel, + TaskRunModel, WorkflowRunModel, ) from skyvern.forge.sdk.db.utils import convert_to_step, convert_to_task, hydrate_action, serialize_proxy_location from skyvern.forge.sdk.models import Step, StepStatus +from skyvern.forge.sdk.schemas.runs import Run from skyvern.forge.sdk.schemas.tasks import OrderBy, SortDirection, Task, TaskStatus from skyvern.forge.sdk.utils.sanitization import sanitize_postgres_text -from skyvern.schemas.runs import ProxyLocationInput +from skyvern.schemas.runs import ProxyLocationInput, RunStatus, RunType from skyvern.schemas.steps import AgentStepOutput from skyvern.webeye.actions.actions import Action @@ -737,3 +739,197 @@ class TasksMixin: ) await session.execute(stmt) await session.commit() + + async def sync_task_run_status( + self, + organization_id: str, + run_id: str, + status: str, + started_at: datetime | None = None, + finished_at: datetime | None = None, + ) -> None: + """Best-effort write-through: propagate status from source table to task_runs. + + Does NOT raise if the task_runs row is missing (race at creation time). + """ + try: + async with self.Session() as session: + vals: dict[str, Any] = {"status": status} + if started_at is not None: + vals["started_at"] = started_at + if finished_at is not None: + vals["finished_at"] = finished_at + stmt = ( + update(TaskRunModel) + .where(TaskRunModel.run_id == run_id) + .where(TaskRunModel.organization_id == organization_id) + .values(**vals) + ) + await session.execute(stmt) + await session.commit() + except Exception: + LOG.warning( + "Best-effort task_run status sync failed", + run_id=run_id, + organization_id=organization_id, + status=status, + exc_info=True, + ) + + @db_operation("create_task_run") + async def create_task_run( + self, + task_run_type: RunType, + organization_id: str, + run_id: str, + title: str | None = None, + url: str | None = None, + url_hash: str | None = None, + status: RunStatus | None = None, + workflow_permanent_id: str | None = None, + parent_workflow_run_id: str | None = None, + debug_session_id: str | None = None, + # script_run, started_at, finished_at are intentionally omitted here — + # they are set via update_task_run() after the run starts/finishes (PRs 2-5). + ) -> Run: + searchable_text = " ".join(filter(None, [title, url])) + async with self.Session() as session: + task_run = TaskRunModel( + task_run_type=task_run_type, + organization_id=organization_id, + run_id=run_id, + title=title, + url=url, + url_hash=url_hash, + status=status, + workflow_permanent_id=workflow_permanent_id, + parent_workflow_run_id=parent_workflow_run_id, + debug_session_id=debug_session_id, + searchable_text=searchable_text or None, + ) + session.add(task_run) + await session.commit() + await session.refresh(task_run) + return Run.model_validate(task_run) + + @db_operation("update_task_run") + async def update_task_run( + self, + organization_id: str, + run_id: str, + title: str | None = None, + url: str | None = None, + url_hash: str | None = None, + status: str | None = None, + started_at: datetime | None = None, + finished_at: datetime | None = None, + ) -> None: + async with self.Session() as session: + task_run = ( + await session.scalars( + select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id) + ) + ).first() + if not task_run: + raise NotFoundError(f"TaskRun {run_id} not found") + + if title is not None: + task_run.title = title + if url is not None: + task_run.url = url + if url_hash is not None: + task_run.url_hash = url_hash + if status is not None: + task_run.status = status + if started_at is not None: + task_run.started_at = started_at + if finished_at is not None: + task_run.finished_at = finished_at + + # Recompute searchable_text when title or url changes + if title is not None or url is not None: + task_run.searchable_text = " ".join(filter(None, [task_run.title, task_run.url])) or None + + await session.commit() + + @db_operation("update_job_run_compute_cost") + async def update_job_run_compute_cost( + self, + organization_id: str, + run_id: str, + instance_type: str | None = None, + vcpu_millicores: int | None = None, + memory_mb: int | None = None, + duration_ms: int | None = None, + compute_cost: float | None = None, + ) -> None: + """Update compute cost metrics for a job run.""" + async with self.Session() as session: + task_run = ( + await session.scalars( + select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id) + ) + ).first() + if not task_run: + LOG.warning( + "TaskRun not found for compute cost update", + run_id=run_id, + organization_id=organization_id, + ) + return + + if instance_type is not None: + task_run.instance_type = instance_type + if vcpu_millicores is not None: + task_run.vcpu_millicores = vcpu_millicores + if memory_mb is not None: + task_run.memory_mb = memory_mb + if duration_ms is not None: + task_run.duration_ms = duration_ms + if compute_cost is not None: + task_run.compute_cost = compute_cost + await session.commit() + + @db_operation("cache_task_run") + async def cache_task_run(self, run_id: str, organization_id: str | None = None) -> Run: + async with self.Session() as session: + task_run = ( + await session.scalars( + select(TaskRunModel).filter_by(organization_id=organization_id).filter_by(run_id=run_id) + ) + ).first() + if task_run: + task_run.cached = True + await session.commit() + await session.refresh(task_run) + return Run.model_validate(task_run) + raise NotFoundError(f"Run {run_id} not found") + + @db_operation("get_cached_task_run") + async def get_cached_task_run( + self, task_run_type: RunType, url_hash: str | None = None, organization_id: str | None = None + ) -> Run | None: + async with self.Session() as session: + query = select(TaskRunModel) + if task_run_type: + query = query.filter_by(task_run_type=task_run_type) + if url_hash: + query = query.filter_by(url_hash=url_hash) + if organization_id: + query = query.filter_by(organization_id=organization_id) + query = query.filter_by(cached=True).order_by(TaskRunModel.created_at.desc()) + task_run = (await session.scalars(query)).first() + return Run.model_validate(task_run) if task_run else None + + @db_operation("get_run") + async def get_run( + self, + run_id: str, + organization_id: str | None = None, + ) -> Run | None: + async with self.Session() as session: + query = select(TaskRunModel).filter_by(run_id=run_id) + if organization_id: + query = query.filter_by(organization_id=organization_id) + task_run = (await session.scalars(query)).first() + return Run.model_validate(task_run) if task_run else None diff --git a/skyvern/forge/sdk/db/mixins/workflow_parameters.py b/skyvern/forge/sdk/db/mixins/workflow_parameters.py index ac474b612..059233d1e 100644 --- a/skyvern/forge/sdk/db/mixins/workflow_parameters.py +++ b/skyvern/forge/sdk/db/mixins/workflow_parameters.py @@ -24,7 +24,6 @@ from skyvern.forge.sdk.db.models import ( OutputParameterModel, TaskGenerationModel, TaskModel, - TaskRunModel, WorkflowCopilotChatMessageModel, WorkflowCopilotChatModel, WorkflowParameterModel, @@ -37,7 +36,6 @@ from skyvern.forge.sdk.db.utils import ( hydrate_action, ) from skyvern.forge.sdk.schemas.ai_suggestions import AISuggestion -from skyvern.forge.sdk.schemas.runs import Run from skyvern.forge.sdk.schemas.task_generations import TaskGeneration from skyvern.forge.sdk.schemas.tasks import Task, TaskStatus from skyvern.forge.sdk.schemas.workflow_copilot import ( @@ -59,7 +57,6 @@ from skyvern.forge.sdk.workflow.models.parameter import ( WorkflowParameter, WorkflowParameterType, ) -from skyvern.schemas.runs import RunType from skyvern.webeye.actions.actions import Action if TYPE_CHECKING: @@ -567,135 +564,3 @@ class WorkflowParametersMixin: actions = (await session.scalars(query)).all() return [Action.model_validate(action) for action in actions] - - @db_operation("create_task_run") - async def create_task_run( - self, - task_run_type: RunType, - organization_id: str, - run_id: str, - title: str | None = None, - url: str | None = None, - url_hash: str | None = None, - ) -> Run: - async with self.Session() as session: - task_run = TaskRunModel( - task_run_type=task_run_type, - organization_id=organization_id, - run_id=run_id, - title=title, - url=url, - url_hash=url_hash, - ) - session.add(task_run) - await session.commit() - await session.refresh(task_run) - return Run.model_validate(task_run) - - @db_operation("update_task_run") - async def update_task_run( - self, - organization_id: str, - run_id: str, - title: str | None = None, - url: str | None = None, - url_hash: str | None = None, - ) -> None: - async with self.Session() as session: - task_run = ( - await session.scalars( - select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id) - ) - ).first() - if not task_run: - raise NotFoundError(f"TaskRun {run_id} not found") - - if title: - task_run.title = title - if url: - task_run.url = url - if url_hash: - task_run.url_hash = url_hash - await session.commit() - - @db_operation("update_job_run_compute_cost") - async def update_job_run_compute_cost( - self, - organization_id: str, - run_id: str, - instance_type: str | None = None, - vcpu_millicores: int | None = None, - memory_mb: int | None = None, - duration_ms: int | None = None, - compute_cost: float | None = None, - ) -> None: - """Update compute cost metrics for a job run.""" - async with self.Session() as session: - task_run = ( - await session.scalars( - select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id) - ) - ).first() - if not task_run: - LOG.warning( - "TaskRun not found for compute cost update", - run_id=run_id, - organization_id=organization_id, - ) - return - - if instance_type is not None: - task_run.instance_type = instance_type - if vcpu_millicores is not None: - task_run.vcpu_millicores = vcpu_millicores - if memory_mb is not None: - task_run.memory_mb = memory_mb - if duration_ms is not None: - task_run.duration_ms = duration_ms - if compute_cost is not None: - task_run.compute_cost = compute_cost - await session.commit() - - @db_operation("cache_task_run") - async def cache_task_run(self, run_id: str, organization_id: str | None = None) -> Run: - async with self.Session() as session: - task_run = ( - await session.scalars( - select(TaskRunModel).filter_by(organization_id=organization_id).filter_by(run_id=run_id) - ) - ).first() - if task_run: - task_run.cached = True - await session.commit() - await session.refresh(task_run) - return Run.model_validate(task_run) - raise NotFoundError(f"Run {run_id} not found") - - @db_operation("get_cached_task_run") - async def get_cached_task_run( - self, task_run_type: RunType, url_hash: str | None = None, organization_id: str | None = None - ) -> Run | None: - async with self.Session() as session: - query = select(TaskRunModel) - if task_run_type: - query = query.filter_by(task_run_type=task_run_type) - if url_hash: - query = query.filter_by(url_hash=url_hash) - if organization_id: - query = query.filter_by(organization_id=organization_id) - query = query.filter_by(cached=True).order_by(TaskRunModel.created_at.desc()) - task_run = (await session.scalars(query)).first() - return Run.model_validate(task_run) if task_run else None - - @db_operation("get_run") - async def get_run( - self, - run_id: str, - organization_id: str | None = None, - ) -> Run | None: - async with self.Session() as session: - query = select(TaskRunModel).filter_by(run_id=run_id) - if organization_id: - query = query.filter_by(organization_id=organization_id) - task_run = (await session.scalars(query)).first() - return Run.model_validate(task_run) if task_run else None diff --git a/skyvern/forge/sdk/db/mixins/workflow_runs.py b/skyvern/forge/sdk/db/mixins/workflow_runs.py index b586d8bcc..c5b646d7a 100644 --- a/skyvern/forge/sdk/db/mixins/workflow_runs.py +++ b/skyvern/forge/sdk/db/mixins/workflow_runs.py @@ -13,6 +13,7 @@ from skyvern.forge.sdk.db.exceptions import NotFoundError from skyvern.forge.sdk.db.mixins.base import read_retry from skyvern.forge.sdk.db.models import ( TaskModel, + TaskRunModel, WorkflowModel, WorkflowParameterModel, WorkflowRunBlockModel, @@ -383,6 +384,36 @@ class WorkflowRunsMixin: return runs[lower:upper] + @read_retry() + async def get_all_runs_v2( + self, + organization_id: str, + page: int = 1, + page_size: int = 10, + status: list[str] | None = None, + search_key: str | None = None, + ) -> list[TaskRunModel]: + async with self.Session() as session: + query = ( + select(TaskRunModel) + .filter(TaskRunModel.organization_id == organization_id) + .filter(TaskRunModel.status.isnot(None)) + .filter(TaskRunModel.parent_workflow_run_id.is_(None)) + .filter(TaskRunModel.debug_session_id.is_(None)) + ) + + if status: + query = query.filter(TaskRunModel.status.in_(status)) + + if search_key: + query = query.filter(TaskRunModel.searchable_text.icontains(search_key, autoescape=True)) + + offset = (page - 1) * page_size + query = query.order_by(TaskRunModel.created_at.desc()).offset(offset).limit(page_size) + + result = await session.scalars(query) + return list(result.all()) + @read_retry() @db_operation("get_workflow_run", log_errors=False) async def get_workflow_run( diff --git a/skyvern/forge/sdk/db/models.py b/skyvern/forge/sdk/db/models.py index 1434623f1..a87663a97 100644 --- a/skyvern/forge/sdk/db/models.py +++ b/skyvern/forge/sdk/db/models.py @@ -12,6 +12,7 @@ from sqlalchemy import ( Integer, Numeric, String, + Text, UnicodeText, UniqueConstraint, desc, @@ -65,6 +66,7 @@ from skyvern.forge.sdk.db.id import ( generate_workflow_script_id, generate_workflow_template_id, ) +from skyvern.forge.sdk.schemas.runs import TERMINAL_STATUSES from skyvern.forge.sdk.schemas.task_v2 import ThoughtType @@ -966,6 +968,35 @@ class TaskRunModel(Base): Index("task_run_org_url_index", "organization_id", "url_hash", "cached"), Index("task_run_org_run_id_index", "organization_id", "run_id"), Index("ix_task_runs_org_created_at", "organization_id", "created_at"), + Index( + "ix_task_runs_org_toplevel_created", + "organization_id", + desc("created_at"), + postgresql_using="btree", + postgresql_where=text("parent_workflow_run_id IS NULL AND debug_session_id IS NULL AND status IS NOT NULL"), + ), + Index( + "ix_task_runs_org_status_created", + "organization_id", + "status", + desc("created_at"), + postgresql_using="btree", + ), + Index( + "ix_task_runs_searchable_text_gin", + "searchable_text", + postgresql_using="gin", + postgresql_ops={"searchable_text": "gin_trgm_ops"}, + ), + Index( + "ix_task_runs_nonterminal", + "run_id", + "task_run_type", + postgresql_where=sqlalchemy.or_( + sqlalchemy.column("status").is_(None), + ~sqlalchemy.column("status").in_(TERMINAL_STATUSES), + ), + ), ) task_run_id = Column(String, primary_key=True, default=generate_task_run_id) @@ -976,6 +1007,17 @@ class TaskRunModel(Base): url = Column(String, nullable=True) url_hash = Column(String, nullable=True) cached = Column(Boolean, nullable=False, default=False) + # Run history fields + # status is an open-ended str (not an enum) because task_runs covers multiple + # run types (task, workflow, observer) each with its own status set. + status = Column(String, nullable=True) + started_at = Column(DateTime, nullable=True) + finished_at = Column(DateTime, nullable=True) + workflow_permanent_id = Column(String, nullable=True) + script_run = Column(JSON, nullable=True) + parent_workflow_run_id = Column(String, nullable=True) + debug_session_id = Column(String, nullable=True) + searchable_text = Column(Text, nullable=True) # Compute cost tracking fields instance_type = Column(String, nullable=True) vcpu_millicores = Column(Integer, nullable=True) diff --git a/skyvern/forge/sdk/db/repositories/organizations.py b/skyvern/forge/sdk/db/repositories/organizations.py index df9168599..4691416cf 100644 --- a/skyvern/forge/sdk/db/repositories/organizations.py +++ b/skyvern/forge/sdk/db/repositories/organizations.py @@ -129,9 +129,11 @@ class OrganizationsRepository(BaseRepository): max_steps_per_run: int | None = None, max_retries_per_step: int | None = None, domain: str | None = None, + organization_id: str | None = None, ) -> Organization: async with self.Session() as session: org = OrganizationModel( + organization_id=organization_id, organization_name=organization_name, webhook_callback_url=webhook_callback_url, max_steps_per_run=max_steps_per_run, diff --git a/skyvern/forge/sdk/db/repositories/tasks.py b/skyvern/forge/sdk/db/repositories/tasks.py index 889aa0ca8..6e54aa086 100644 --- a/skyvern/forge/sdk/db/repositories/tasks.py +++ b/skyvern/forge/sdk/db/repositories/tasks.py @@ -1,5 +1,6 @@ from __future__ import annotations +import asyncio from datetime import datetime, timedelta, timezone from typing import Any, Sequence @@ -15,13 +16,15 @@ from skyvern.forge.sdk.db.models import ( ActionModel, StepModel, TaskModel, + TaskRunModel, WorkflowRunModel, ) from skyvern.forge.sdk.db.utils import convert_to_step, convert_to_task, hydrate_action, serialize_proxy_location from skyvern.forge.sdk.models import Step, StepStatus +from skyvern.forge.sdk.schemas.runs import Run from skyvern.forge.sdk.schemas.tasks import OrderBy, SortDirection, Task, TaskStatus from skyvern.forge.sdk.utils.sanitization import sanitize_postgres_text -from skyvern.schemas.runs import ProxyLocationInput +from skyvern.schemas.runs import ProxyLocationInput, RunStatus, RunType from skyvern.schemas.steps import AgentStepOutput from skyvern.webeye.actions.actions import Action @@ -29,6 +32,8 @@ LOG = structlog.get_logger() class TasksRepository(BaseRepository): + _background_tasks: set[asyncio.Task] = set() # noqa: RUF012 + @db_operation("create_task") async def create_task( self, @@ -470,6 +475,22 @@ class TasksRepository(BaseRepository): updated_task = await self.get_task(task_id, organization_id=organization_id) if not updated_task: raise NotFoundError("Task not found") + + # Best-effort fire-and-forget write-through to task_runs. + # Mirrors the WorkflowService pattern — cron catches any missed syncs. + if status is not None: + bg = asyncio.create_task( + self.sync_task_run_status( + organization_id=updated_task.organization_id or "", + run_id=updated_task.task_id, + status=status.value, + started_at=updated_task.started_at, + finished_at=updated_task.finished_at, + ), + ) + self._background_tasks.add(bg) + bg.add_done_callback(self._background_tasks.discard) + return updated_task else: raise NotFoundError("Task not found") @@ -729,3 +750,197 @@ class TasksRepository(BaseRepository): ) await session.execute(stmt) await session.commit() + + async def sync_task_run_status( + self, + organization_id: str, + run_id: str, + status: str, + started_at: datetime | None = None, + finished_at: datetime | None = None, + ) -> None: + """Best-effort write-through: propagate status from source table to task_runs. + + Does NOT raise if the task_runs row is missing (race at creation time). + """ + try: + async with self.Session() as session: + vals: dict[str, Any] = {"status": status} + if started_at is not None: + vals["started_at"] = started_at + if finished_at is not None: + vals["finished_at"] = finished_at + stmt = ( + update(TaskRunModel) + .where(TaskRunModel.run_id == run_id) + .where(TaskRunModel.organization_id == organization_id) + .values(**vals) + ) + await session.execute(stmt) + await session.commit() + except Exception: + LOG.warning( + "Best-effort task_run status sync failed", + run_id=run_id, + organization_id=organization_id, + status=status, + exc_info=True, + ) + + @db_operation("create_task_run") + async def create_task_run( + self, + task_run_type: RunType, + organization_id: str, + run_id: str, + title: str | None = None, + url: str | None = None, + url_hash: str | None = None, + status: RunStatus | None = None, + workflow_permanent_id: str | None = None, + parent_workflow_run_id: str | None = None, + debug_session_id: str | None = None, + # script_run, started_at, finished_at are intentionally omitted here — + # they are set via update_task_run() after the run starts/finishes (PRs 2-5). + ) -> Run: + searchable_text = " ".join(filter(None, [title, url])) + async with self.Session() as session: + task_run = TaskRunModel( + task_run_type=task_run_type, + organization_id=organization_id, + run_id=run_id, + title=title, + url=url, + url_hash=url_hash, + status=status, + workflow_permanent_id=workflow_permanent_id, + parent_workflow_run_id=parent_workflow_run_id, + debug_session_id=debug_session_id, + searchable_text=searchable_text or None, + ) + session.add(task_run) + await session.commit() + await session.refresh(task_run) + return Run.model_validate(task_run) + + @db_operation("update_task_run") + async def update_task_run( + self, + organization_id: str, + run_id: str, + title: str | None = None, + url: str | None = None, + url_hash: str | None = None, + status: str | None = None, + started_at: datetime | None = None, + finished_at: datetime | None = None, + ) -> None: + async with self.Session() as session: + task_run = ( + await session.scalars( + select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id) + ) + ).first() + if not task_run: + raise NotFoundError(f"TaskRun {run_id} not found") + + if title is not None: + task_run.title = title + if url is not None: + task_run.url = url + if url_hash is not None: + task_run.url_hash = url_hash + if status is not None: + task_run.status = status + if started_at is not None: + task_run.started_at = started_at + if finished_at is not None: + task_run.finished_at = finished_at + + # Recompute searchable_text when title or url changes + if title is not None or url is not None: + task_run.searchable_text = " ".join(filter(None, [task_run.title, task_run.url])) or None + + await session.commit() + + @db_operation("update_job_run_compute_cost") + async def update_job_run_compute_cost( + self, + organization_id: str, + run_id: str, + instance_type: str | None = None, + vcpu_millicores: int | None = None, + memory_mb: int | None = None, + duration_ms: int | None = None, + compute_cost: float | None = None, + ) -> None: + """Update compute cost metrics for a job run.""" + async with self.Session() as session: + task_run = ( + await session.scalars( + select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id) + ) + ).first() + if not task_run: + LOG.warning( + "TaskRun not found for compute cost update", + run_id=run_id, + organization_id=organization_id, + ) + return + + if instance_type is not None: + task_run.instance_type = instance_type + if vcpu_millicores is not None: + task_run.vcpu_millicores = vcpu_millicores + if memory_mb is not None: + task_run.memory_mb = memory_mb + if duration_ms is not None: + task_run.duration_ms = duration_ms + if compute_cost is not None: + task_run.compute_cost = compute_cost + await session.commit() + + @db_operation("cache_task_run") + async def cache_task_run(self, run_id: str, organization_id: str | None = None) -> Run: + async with self.Session() as session: + task_run = ( + await session.scalars( + select(TaskRunModel).filter_by(organization_id=organization_id).filter_by(run_id=run_id) + ) + ).first() + if task_run: + task_run.cached = True + await session.commit() + await session.refresh(task_run) + return Run.model_validate(task_run) + raise NotFoundError(f"Run {run_id} not found") + + @db_operation("get_cached_task_run") + async def get_cached_task_run( + self, task_run_type: RunType, url_hash: str | None = None, organization_id: str | None = None + ) -> Run | None: + async with self.Session() as session: + query = select(TaskRunModel) + if task_run_type: + query = query.filter_by(task_run_type=task_run_type) + if url_hash: + query = query.filter_by(url_hash=url_hash) + if organization_id: + query = query.filter_by(organization_id=organization_id) + query = query.filter_by(cached=True).order_by(TaskRunModel.created_at.desc()) + task_run = (await session.scalars(query)).first() + return Run.model_validate(task_run) if task_run else None + + @db_operation("get_run") + async def get_run( + self, + run_id: str, + organization_id: str | None = None, + ) -> Run | None: + async with self.Session() as session: + query = select(TaskRunModel).filter_by(run_id=run_id) + if organization_id: + query = query.filter_by(organization_id=organization_id) + task_run = (await session.scalars(query)).first() + return Run.model_validate(task_run) if task_run else None diff --git a/skyvern/forge/sdk/db/repositories/workflow_parameters.py b/skyvern/forge/sdk/db/repositories/workflow_parameters.py index 9eb5fe80d..b48a43fc4 100644 --- a/skyvern/forge/sdk/db/repositories/workflow_parameters.py +++ b/skyvern/forge/sdk/db/repositories/workflow_parameters.py @@ -26,7 +26,6 @@ from skyvern.forge.sdk.db.models import ( OutputParameterModel, TaskGenerationModel, TaskModel, - TaskRunModel, WorkflowCopilotChatMessageModel, WorkflowCopilotChatModel, WorkflowParameterModel, @@ -39,7 +38,6 @@ from skyvern.forge.sdk.db.utils import ( hydrate_action, ) from skyvern.forge.sdk.schemas.ai_suggestions import AISuggestion -from skyvern.forge.sdk.schemas.runs import Run from skyvern.forge.sdk.schemas.task_generations import TaskGeneration from skyvern.forge.sdk.schemas.tasks import Task, TaskStatus from skyvern.forge.sdk.schemas.workflow_copilot import ( @@ -61,7 +59,6 @@ from skyvern.forge.sdk.workflow.models.parameter import ( WorkflowParameter, WorkflowParameterType, ) -from skyvern.schemas.runs import RunType from skyvern.webeye.actions.actions import Action LOG = structlog.get_logger() @@ -564,135 +561,3 @@ class WorkflowParametersRepository(BaseRepository): actions = (await session.scalars(query)).all() return [Action.model_validate(action) for action in actions] - - @db_operation("create_task_run") - async def create_task_run( - self, - task_run_type: RunType, - organization_id: str, - run_id: str, - title: str | None = None, - url: str | None = None, - url_hash: str | None = None, - ) -> Run: - async with self.Session() as session: - task_run = TaskRunModel( - task_run_type=task_run_type, - organization_id=organization_id, - run_id=run_id, - title=title, - url=url, - url_hash=url_hash, - ) - session.add(task_run) - await session.commit() - await session.refresh(task_run) - return Run.model_validate(task_run) - - @db_operation("update_task_run") - async def update_task_run( - self, - organization_id: str, - run_id: str, - title: str | None = None, - url: str | None = None, - url_hash: str | None = None, - ) -> None: - async with self.Session() as session: - task_run = ( - await session.scalars( - select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id) - ) - ).first() - if not task_run: - raise NotFoundError(f"TaskRun {run_id} not found") - - if title: - task_run.title = title - if url: - task_run.url = url - if url_hash: - task_run.url_hash = url_hash - await session.commit() - - @db_operation("update_job_run_compute_cost") - async def update_job_run_compute_cost( - self, - organization_id: str, - run_id: str, - instance_type: str | None = None, - vcpu_millicores: int | None = None, - memory_mb: int | None = None, - duration_ms: int | None = None, - compute_cost: float | None = None, - ) -> None: - """Update compute cost metrics for a job run.""" - async with self.Session() as session: - task_run = ( - await session.scalars( - select(TaskRunModel).filter_by(run_id=run_id).filter_by(organization_id=organization_id) - ) - ).first() - if not task_run: - LOG.warning( - "TaskRun not found for compute cost update", - run_id=run_id, - organization_id=organization_id, - ) - return - - if instance_type is not None: - task_run.instance_type = instance_type - if vcpu_millicores is not None: - task_run.vcpu_millicores = vcpu_millicores - if memory_mb is not None: - task_run.memory_mb = memory_mb - if duration_ms is not None: - task_run.duration_ms = duration_ms - if compute_cost is not None: - task_run.compute_cost = compute_cost - await session.commit() - - @db_operation("cache_task_run") - async def cache_task_run(self, run_id: str, organization_id: str | None = None) -> Run: - async with self.Session() as session: - task_run = ( - await session.scalars( - select(TaskRunModel).filter_by(organization_id=organization_id).filter_by(run_id=run_id) - ) - ).first() - if task_run: - task_run.cached = True - await session.commit() - await session.refresh(task_run) - return Run.model_validate(task_run) - raise NotFoundError(f"Run {run_id} not found") - - @db_operation("get_cached_task_run") - async def get_cached_task_run( - self, task_run_type: RunType, url_hash: str | None = None, organization_id: str | None = None - ) -> Run | None: - async with self.Session() as session: - query = select(TaskRunModel) - if task_run_type: - query = query.filter_by(task_run_type=task_run_type) - if url_hash: - query = query.filter_by(url_hash=url_hash) - if organization_id: - query = query.filter_by(organization_id=organization_id) - query = query.filter_by(cached=True).order_by(TaskRunModel.created_at.desc()) - task_run = (await session.scalars(query)).first() - return Run.model_validate(task_run) if task_run else None - - @db_operation("get_run") - async def get_run( - self, - run_id: str, - organization_id: str | None = None, - ) -> Run | None: - async with self.Session() as session: - query = select(TaskRunModel).filter_by(run_id=run_id) - if organization_id: - query = query.filter_by(organization_id=organization_id) - task_run = (await session.scalars(query)).first() - return Run.model_validate(task_run) if task_run else None diff --git a/skyvern/forge/sdk/db/repositories/workflow_runs.py b/skyvern/forge/sdk/db/repositories/workflow_runs.py index 604f5b988..a8acb8ac3 100644 --- a/skyvern/forge/sdk/db/repositories/workflow_runs.py +++ b/skyvern/forge/sdk/db/repositories/workflow_runs.py @@ -21,6 +21,7 @@ if TYPE_CHECKING: from skyvern.forge.sdk.db._sentinels import _UNSET from skyvern.forge.sdk.db.models import ( TaskModel, + TaskRunModel, WorkflowModel, WorkflowParameterModel, WorkflowRunBlockModel, @@ -392,6 +393,36 @@ class WorkflowRunsRepository(BaseRepository): return runs[lower:upper] + @read_retry() + async def get_all_runs_v2( + self, + organization_id: str, + page: int = 1, + page_size: int = 10, + status: list[str] | None = None, + search_key: str | None = None, + ) -> list[TaskRunModel]: + async with self.Session() as session: + query = ( + select(TaskRunModel) + .filter(TaskRunModel.organization_id == organization_id) + .filter(TaskRunModel.status.isnot(None)) + .filter(TaskRunModel.parent_workflow_run_id.is_(None)) + .filter(TaskRunModel.debug_session_id.is_(None)) + ) + + if status: + query = query.filter(TaskRunModel.status.in_(status)) + + if search_key: + query = query.filter(TaskRunModel.searchable_text.icontains(search_key, autoescape=True)) + + offset = (page - 1) * page_size + query = query.order_by(TaskRunModel.created_at.desc()).offset(offset).limit(page_size) + + result = await session.scalars(query) + return list(result.all()) + @read_retry() @db_operation("get_workflow_run", log_errors=False) async def get_workflow_run( diff --git a/skyvern/forge/sdk/routes/agent_protocol.py b/skyvern/forge/sdk/routes/agent_protocol.py index 55c3783eb..4620fddf0 100644 --- a/skyvern/forge/sdk/routes/agent_protocol.py +++ b/skyvern/forge/sdk/routes/agent_protocol.py @@ -113,7 +113,9 @@ from skyvern.schemas.runs import ( BlockRunResponse, RunEngine, RunResponse, + RunStatus, RunType, + TaskRunListItem, TaskRunRequest, TaskRunResponse, UploadFileResponse, @@ -1082,7 +1084,6 @@ async def delete_workflow( "/folders", response_model=Folder, tags=["Workflow Folders"], - include_in_schema=False, openapi_extra={ "x-fern-sdk-method-name": "create_folder", }, @@ -1125,7 +1126,6 @@ async def create_folder( "/folders/{folder_id}", response_model=Folder, tags=["Workflow Folders"], - include_in_schema=False, openapi_extra={ "x-fern-sdk-method-name": "get_folder", }, @@ -1170,7 +1170,6 @@ async def get_folder( "/folders", response_model=list[Folder], tags=["Workflow Folders"], - include_in_schema=False, openapi_extra={ "x-fern-sdk-method-name": "get_folders", }, @@ -1228,7 +1227,6 @@ async def get_folders( "/folders/{folder_id}", response_model=Folder, tags=["Workflow Folders"], - include_in_schema=False, openapi_extra={ "x-fern-sdk-method-name": "update_folder", }, @@ -1275,7 +1273,6 @@ async def update_folder( @base_router.delete( "/folders/{folder_id}", tags=["Workflow Folders"], - include_in_schema=False, openapi_extra={ "x-fern-sdk-method-name": "delete_folder", }, @@ -1312,7 +1309,6 @@ async def delete_folder( "/workflows/{workflow_permanent_id}/folder", response_model=Workflow, tags=["Workflow Folders"], - include_in_schema=False, openapi_extra={ "x-fern-sdk-method-name": "update_workflow_folder", }, @@ -2147,6 +2143,48 @@ async def get_runs( return ORJSONResponse([run.model_dump() for run in runs]) +# NOTE: v2 returns TaskRunListItem from the unified task_runs table, +# replacing the v1 response type (list[WorkflowRun | Task]) which +# merged two separate queries. The v1 endpoint is preserved for +# backwards compatibility until clients migrate. +@base_router.get( + "/runs", + tags=["agent"], + response_model=list[TaskRunListItem], + openapi_extra={ + "x-fern-sdk-method-name": "get_runs_v2", + }, +) +@base_router.get( + "/runs/", + response_model=list[TaskRunListItem], + include_in_schema=False, +) +async def get_runs_v2( + current_org: Organization = Depends(org_auth_service.get_current_org), + page: int = Query(1, ge=1, le=100), + page_size: int = Query(10, ge=1, le=100), + status: Annotated[list[RunStatus] | None, Query()] = None, + search_key: str | None = Query( + None, + min_length=3, + description="Case-insensitive substring search (min 3 chars for trigram index).", + examples=["login_url", "wr_abc123"], + ), +) -> Response: + analytics.capture("skyvern-oss-agent-runs-v2-get") + + rows = await app.DATABASE.get_all_runs_v2( + current_org.organization_id, + page=page, + page_size=page_size, + status=[s.value for s in status] if status else None, + search_key=search_key, + ) + items = [TaskRunListItem.model_validate(row) for row in rows] + return ORJSONResponse([item.model_dump(mode="json") for item in items]) + + @legacy_base_router.get( "/tasks/{task_id}/steps", tags=["agent"], diff --git a/skyvern/forge/sdk/schemas/runs.py b/skyvern/forge/sdk/schemas/runs.py index dab90a499..5e284cb99 100644 --- a/skyvern/forge/sdk/schemas/runs.py +++ b/skyvern/forge/sdk/schemas/runs.py @@ -1,8 +1,8 @@ from datetime import datetime -from pydantic import BaseModel, ConfigDict +from pydantic import BaseModel, ConfigDict, Field -from skyvern.schemas.runs import RunType +from skyvern.schemas.runs import TERMINAL_STATUSES, RunType # noqa: F401 class Run(BaseModel): @@ -15,6 +15,18 @@ class Run(BaseModel): title: str | None = None url: str | None = None cached: bool = False + # Run history fields + status: str | None = None + started_at: datetime | None = None + finished_at: datetime | None = None + workflow_permanent_id: str | None = None + # dict when script execution metadata is present (e.g. {"ai_fallback_triggered": True}), + # bool (True/False) when only the ran-as-script flag is needed, None when not a script run. + script_run: dict | bool | None = None + parent_workflow_run_id: str | None = None + debug_session_id: str | None = None + # Internal denormalized column for trigram search — excluded from serialization. + searchable_text: str | None = Field(default=None, exclude=True) # Compute cost tracking fields instance_type: str | None = None vcpu_millicores: int | None = None diff --git a/skyvern/forge/sdk/services/local_org_auth_token_service.py b/skyvern/forge/sdk/services/local_org_auth_token_service.py index 808e43115..19313b720 100644 --- a/skyvern/forge/sdk/services/local_org_auth_token_service.py +++ b/skyvern/forge/sdk/services/local_org_auth_token_service.py @@ -1,13 +1,19 @@ import os +import time from pathlib import Path +import jwt import structlog from dotenv import set_key +from jwt.exceptions import PyJWTError +from pydantic import ValidationError from skyvern.config import settings from skyvern.forge import app from skyvern.forge.sdk.core import security -from skyvern.forge.sdk.schemas.organizations import Organization, OrganizationAuthTokenType +from skyvern.forge.sdk.db.enums import OrganizationAuthTokenType +from skyvern.forge.sdk.models import TokenPayload +from skyvern.forge.sdk.schemas.organizations import Organization from skyvern.forge.sdk.services.org_auth_token_service import API_KEY_LIFETIME from skyvern.utils.env_paths import resolve_backend_env_path, resolve_frontend_env_path @@ -42,6 +48,89 @@ async def ensure_local_org() -> Organization: ) +async def ensure_local_org_with_id(organization_id: str) -> Organization: + """Ensure the local org exists, preferring the caller-provided ID on first creation. + + If a local org already exists for the shared local domain, this returns that + org unchanged even when its organization_id differs from the requested one. + """ + organization = await app.DATABASE.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + if organization: + return organization + + return await app.DATABASE.create_organization( + organization_id=organization_id, + organization_name=SKYVERN_LOCAL_ORG, + domain=SKYVERN_LOCAL_DOMAIN, + max_steps_per_run=10, + max_retries_per_step=3, + ) + + +def _decode_local_api_key_payload(api_key: str) -> TokenPayload | None: + try: + payload = jwt.decode( + api_key, + settings.SECRET_KEY, + algorithms=[settings.SIGNATURE_ALGORITHM], + options={"verify_exp": False}, + ) + return TokenPayload(**payload) + except (PyJWTError, ValidationError): + return None + + +async def ensure_local_api_key(api_key: str) -> tuple[str, str] | None: + """Ensure the provided API key remains usable for the local organization. + + Preserves the externally provided key by syncing it into the database + instead of rewriting env files, but only when it is still a valid local JWT. + Returns ``None`` when the key cannot be preserved and the caller should + regenerate a new local key. + """ + payload = _decode_local_api_key_payload(api_key) + if payload is None: + LOG.warning("Existing local API key is not a valid JWT; regenerating", fingerprint=fingerprint_token(api_key)) + return None + if payload.exp < time.time(): + LOG.warning("Existing local API key is expired; regenerating", fingerprint=fingerprint_token(api_key)) + return None + + organization = await app.DATABASE.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + if organization is None: + organization = await ensure_local_org_with_id(payload.sub) + elif organization.organization_id != payload.sub: + LOG.warning( + "Existing local organization does not match API key subject; regenerating", + existing_organization_id=organization.organization_id, + token_organization_id=payload.sub, + fingerprint=fingerprint_token(api_key), + ) + return None + + org_id = organization.organization_id + existing_token = await app.DATABASE.validate_org_auth_token( + organization_id=org_id, + token_type=OrganizationAuthTokenType.api, + token=api_key, + ) + if existing_token is None: + await app.DATABASE.replace_org_auth_token( + organization_id=org_id, + token_type=OrganizationAuthTokenType.api, + token=api_key, + ) + LOG.info( + "Local API key synced", + organization_id=org_id, + fingerprint=fingerprint_token(api_key), + ) + + settings.SKYVERN_API_KEY = api_key + os.environ["SKYVERN_API_KEY"] = api_key + return api_key, org_id + + async def regenerate_local_api_key() -> tuple[str, str, str, str | None]: """Create a fresh API key for the local organization and persist it to env files. diff --git a/skyvern/forge/sdk/workflow/service.py b/skyvern/forge/sdk/workflow/service.py index 5e423de6f..970d1052d 100644 --- a/skyvern/forge/sdk/workflow/service.py +++ b/skyvern/forge/sdk/workflow/service.py @@ -299,6 +299,9 @@ def _get_workflow_definition_core_data(workflow_definition: WorkflowDefinition) class WorkflowService: + # Prevent GC of fire-and-forget asyncio tasks (e.g. task_run sync). + _background_tasks: set[asyncio.Task] = set() # noqa: RUF012 + @staticmethod def _determine_cache_invalidation( previous_blocks: list[dict[str, Any]], @@ -1011,7 +1014,9 @@ class WorkflowService: browser_session_id=browser_session_id, workflow_run_id=workflow_run_id, ) - failure_reason = f"Failed to begin browser session for workflow run: {str(e)}" + failure_reason = ( + f"Failed to begin browser session for workflow run: {get_user_facing_exception_message(e)}" + ) workflow_run = await self.mark_workflow_run_as_failed( workflow_run_id=workflow_run_id, failure_reason=failure_reason, @@ -3623,8 +3628,51 @@ class WorkflowService: trigger_type=workflow_run.trigger_type, workflow_schedule_id=workflow_run.workflow_schedule_id, ) + # Best-effort fire-and-forget write-through to task_runs table. + # Runs off the hot path so workflow status transitions stay fast. + bg = asyncio.create_task( + self._sync_task_run_from_workflow_run(workflow_run, workflow_run_id, status), + ) + self._background_tasks.add(bg) + bg.add_done_callback(self._background_tasks.discard) + return workflow_run + async def _sync_task_run_from_workflow_run( + self, + workflow_run: WorkflowRun, + workflow_run_id: str, + status: WorkflowRunStatus, + ) -> None: + """Fire-and-forget: propagate workflow_run status to task_runs.""" + try: + await app.DATABASE.sync_task_run_status( + organization_id=workflow_run.organization_id, + run_id=workflow_run_id, + status=status.value, + started_at=workflow_run.started_at, + finished_at=workflow_run.finished_at, + ) + # Also sync task_v2 if this workflow_run backs an observer_cruise + task_v2 = await app.DATABASE.get_task_v2_by_workflow_run_id( + workflow_run_id=workflow_run_id, + organization_id=workflow_run.organization_id, + ) + if task_v2: + await app.DATABASE.sync_task_run_status( + organization_id=workflow_run.organization_id, + run_id=task_v2.observer_cruise_id, + status=status.value, + started_at=workflow_run.started_at, + finished_at=workflow_run.finished_at, + ) + except Exception: + LOG.warning( + "Failed to sync task_run status from workflow_run", + workflow_run_id=workflow_run_id, + exc_info=True, + ) + async def mark_workflow_run_as_completed(self, workflow_run_id: str, run_with: str | None = None) -> WorkflowRun: LOG.info( f"Marking workflow run {workflow_run_id} as completed", diff --git a/skyvern/library/embedded_server_factory.py b/skyvern/library/embedded_server_factory.py index 1562841ca..70ab08a5d 100644 --- a/skyvern/library/embedded_server_factory.py +++ b/skyvern/library/embedded_server_factory.py @@ -1,3 +1,4 @@ +import copy import os import tempfile from datetime import timedelta @@ -15,6 +16,57 @@ from skyvern.forge.sdk.api.llm.models import LLMConfig, LLMRouterConfig _EMBEDDED_ACTIVE: bool = False _BLOCKED_EMBEDDED_SETTINGS = frozenset({"OTEL_ENABLED", "ENABLE_CLEANUP_CRON"}) +_SQLITE_OVERRIDE_VALUES: dict[str, Any] = { + "DATABASE_STRING": "sqlite+aiosqlite:///:memory:", + "DATABASE_REPLICA_STRING": None, + "ADDITIONAL_MODULES": [], + "OTEL_ENABLED": False, + "ENABLE_CLEANUP_CRON": False, +} +_BOOTSTRAP_RUNTIME_SETTINGS = frozenset({"LLM_KEY", "BROWSER_LOGS_ENABLED", "SKYVERN_API_KEY"}) + +# Keys mutated by _apply_sqlite_overrides and bootstrap-time request setup. +# SQLite override keys are derived from the actual override mapping so that +# snapshot coverage can't silently drift when new overrides are added. +# OTEL_ENABLED and ENABLE_CLEANUP_CRON overlap with _BLOCKED_EMBEDDED_SETTINGS — +# snapshotting them is a defensive no-op (blocked keys can't be mutated via +# settings_overrides, but _apply_sqlite_overrides sets them directly). +_SETTINGS_SNAPSHOT_KEYS = frozenset(_SQLITE_OVERRIDE_VALUES) | _BLOCKED_EMBEDDED_SETTINGS | _BOOTSTRAP_RUNTIME_SETTINGS + + +def _snapshot_settings() -> dict[str, dict[str, Any]]: + """Capture current values of mutable settings keys across all targets.""" + from skyvern.forge.sdk.settings_manager import SettingsManager # noqa: PLC0415 + + snapshots: dict[str, dict[str, Any]] = {} + targets = {"settings": settings} + mgr = SettingsManager.get_settings() + if mgr is not settings: + targets["mgr"] = mgr + + for label, target in targets.items(): + snap: dict[str, Any] = {} + for key in _SETTINGS_SNAPSHOT_KEYS: + if hasattr(target, key): + val = getattr(target, key) + snap[key] = copy.deepcopy(val) if isinstance(val, (list, dict)) else val + snapshots[label] = snap + return snapshots + + +def _restore_settings(snapshots: dict[str, dict[str, Any]]) -> None: + """Restore settings from a snapshot taken by _snapshot_settings.""" + from skyvern.forge.sdk.settings_manager import SettingsManager # noqa: PLC0415 + + targets = {"settings": settings} + mgr = SettingsManager.get_settings() + if mgr is not settings: + targets["mgr"] = mgr + + for label, target in targets.items(): + if label in snapshots: + for key, val in snapshots[label].items(): + setattr(target, key, val) def create_embedded_server( @@ -39,9 +91,28 @@ def create_embedded_server( async def handle_async_request(self, request: httpx.Request) -> httpx.Response: if self._transport is None: + snapshots = _snapshot_settings() + # Capture the current forge app instance so we can restore it + # if bootstrap fails partway through (after create_api_app sets + # a new instance but before bootstrap completes). + from skyvern.forge import app as forge_app_holder # noqa: PLC0415 + + prev_app_inst = object.__getattribute__(forge_app_holder, "_inst") # type: ignore[arg-type] try: await self._bootstrap(request) - except Exception: + except Exception as exc: + import structlog # noqa: PLC0415 + + structlog.get_logger().exception( + "Embedded bootstrap failed; restoring previous process state", + error_type=type(exc).__name__, + ) + # Restore settings so a retry (or subsequent tests in the + # same process) sees the original values, not half-applied + # SQLite overrides. + _restore_settings(snapshots) + # Restore the previous forge app instance + forge_app_holder.set_app(prev_app_inst) # type: ignore[attr-defined] # Reset the single-client guard so the user can retry # after fixing the issue (e.g., missing API key). global _EMBEDDED_ACTIVE @@ -82,7 +153,9 @@ def create_embedded_server( raise ValueError(f"Invalid setting: {key}") if not use_in_memory_db: - self._api_key = os.getenv("SKYVERN_API_KEY") + self._api_key = ( + settings_overrides.get("SKYVERN_API_KEY") if settings_overrides is not None else None + ) or os.getenv("SKYVERN_API_KEY") if not self._api_key: raise ValueError("SKYVERN_API_KEY is not set. Provide api_key or set SKYVERN_API_KEY in .env file.") @@ -199,9 +272,7 @@ def _apply_sqlite_overrides() -> None: if mgr_settings is not settings: targets.append(mgr_settings) for target in targets: - target.DATABASE_STRING = "sqlite+aiosqlite:///:memory:" - target.DATABASE_REPLICA_STRING = None - target.ADDITIONAL_MODULES = [] - target.OTEL_ENABLED = False - if hasattr(target, "ENABLE_CLEANUP_CRON"): - target.ENABLE_CLEANUP_CRON = False + for key, value in _SQLITE_OVERRIDE_VALUES.items(): + if hasattr(target, key): + copied = copy.deepcopy(value) if isinstance(value, (list, dict)) else value + setattr(target, key, copied) diff --git a/skyvern/library/skyvern.py b/skyvern/library/skyvern.py index 1c80e6b7a..50fc89c23 100644 --- a/skyvern/library/skyvern.py +++ b/skyvern/library/skyvern.py @@ -129,7 +129,7 @@ class Skyvern(AsyncSkyvern): *, llm_config: LLMRouterConfig | LLMConfig | None = None, settings: dict[str, Any] | None = None, - use_in_memory_db: bool = False, + use_in_memory_db: bool | None = None, ) -> "Skyvern": """Local/embedded mode: Run Skyvern locally in-process. @@ -139,23 +139,36 @@ class Skyvern(AsyncSkyvern): overriding the LLM_KEY setting from your .env file. If not provided, uses the LLM configured via LLM_KEY in your .env file. - Example 1 - Using .env configuration (simplest, recommended): + Example 1 - Zero-config (simplest, recommended): ```python from skyvern import Skyvern - # Uses LLM_KEY and other settings from your .env file - # Created by running `skyvern quickstart` + # Auto-detects: uses .env/Postgres if configured, else in-memory SQLite. skyvern = Skyvern.local() ``` - Example 2 - Zero-config with in-memory database (no Postgres needed): + Example 2 - Force in-memory SQLite (no .env needed): ```python from skyvern import Skyvern skyvern = Skyvern.local(use_in_memory_db=True) ``` - Example 3 - Custom LLM with environment variables: + Example 3 - Force persistent mode: + ```python + from skyvern import Skyvern + + # Works with env vars, .env, or explicit settings overrides. + skyvern = Skyvern.local( + use_in_memory_db=False, + settings={ + "DATABASE_STRING": "postgresql+psycopg://skyvern@localhost/skyvern", + "SKYVERN_API_KEY": "sk-...", + }, + ) + ``` + + Example 4 - Custom LLM with environment variables: ```python from skyvern import Skyvern from skyvern.forge.sdk.api.llm.models import LLMConfig @@ -173,16 +186,22 @@ class Skyvern(AsyncSkyvern): settings: Optional dictionary of Skyvern settings to override. These override the corresponding settings from your .env file. Example: {"MAX_STEPS_PER_RUN": 100, "BROWSER_TYPE": "chromium-headful"} - use_in_memory_db: If True, use SQLite in-memory instead of PostgreSQL. - No .env file or running Postgres instance required. Defaults to False. + use_in_memory_db: Controls the database backend for embedded mode. - Zero-config mode supports: + - None (default): Auto-detect. If DATABASE_STRING is set in env, .env, + or settings overrides, use persistent mode (Postgres). Otherwise use + in-memory SQLite for zero-config operation. + - True: Force in-memory SQLite. No .env or Postgres required. + - False: Force persistent mode. Requires DATABASE_STRING and + SKYVERN_API_KEY from env, .env, or settings overrides. + + In-memory mode supports: - run_task(), extract(), click(), navigate() — full browser automation - Workflow CRUD (create, list, search, get, run) - Artifacts saved to local temp directory (file:// URIs) - Any LLM provider supported by litellm - Not supported in zero-config mode (requires Skyvern Cloud or Postgres): + Not supported in in-memory mode (requires Skyvern Cloud or Postgres): - Workflow scheduling (requires persistent database) - Cloud browser sessions (requires S3/Azure storage) - Rate limiting (cloud-only) @@ -194,16 +213,38 @@ class Skyvern(AsyncSkyvern): Returns: Skyvern: A Skyvern instance running in local/embedded mode. """ + from dotenv import dotenv_values # noqa: PLC0415 + from skyvern.library.embedded_server_factory import create_embedded_server # noqa: PLC0415 + from skyvern.utils.env_paths import resolve_backend_env_path # noqa: PLC0415 + + # Auto-detect mode when use_in_memory_db is not explicitly set. + # If DATABASE_STRING is configured anywhere, honor it (persistent mode). + # Otherwise fall back to zero-config in-memory SQLite. + if use_in_memory_db is None: + env_path = resolve_backend_env_path() + dotenv_config = dotenv_values(env_path) if env_path.exists() else {} + settings_overrides = settings or {} + explicit_db = ( + settings_overrides.get("DATABASE_STRING") + or os.environ.get("DATABASE_STRING") + or dotenv_config.get("DATABASE_STRING") + ) + use_in_memory_db = not bool(explicit_db) if not use_in_memory_db: - if not os.path.exists(".env"): - raise ValueError("Please run `skyvern quickstart` to set up your local Skyvern environment") + env_path = resolve_backend_env_path() + if env_path.exists(): + load_dotenv(env_path) - load_dotenv(".env") - api_key = os.getenv("SKYVERN_API_KEY") + settings_overrides = settings or {} + api_key = settings_overrides.get("SKYVERN_API_KEY") or os.getenv("SKYVERN_API_KEY") if not api_key: - raise ValueError("SKYVERN_API_KEY is not set. Provide api_key or set SKYVERN_API_KEY in .env file.") + raise ValueError( + "Persistent local mode requires SKYVERN_API_KEY. " + "Set it in env/.env, pass settings={'SKYVERN_API_KEY': ...}, " + "or use use_in_memory_db=True for ephemeral SQLite." + ) obj = cls.__new__(cls) diff --git a/skyvern/schemas/runs.py b/skyvern/schemas/runs.py index a75f77424..649091888 100644 --- a/skyvern/schemas/runs.py +++ b/skyvern/schemas/runs.py @@ -5,7 +5,7 @@ from enum import StrEnum from typing import Annotated, Any, Literal, Union from zoneinfo import ZoneInfo -from pydantic import BaseModel, Field, field_validator +from pydantic import BaseModel, ConfigDict, Field, field_validator from skyvern.forge.sdk.schemas.files import FileInfo from skyvern.forge.sdk.workflow.models.validators import normalize_run_with @@ -362,7 +362,13 @@ class RunStatus(StrEnum): canceled = "canceled" def is_final(self) -> bool: - return self in [self.failed, self.terminated, self.canceled, self.timed_out, self.completed] + return self.value in TERMINAL_STATUSES + + +# Statuses that are final — once a row reaches one of these, it never changes. +# Single source of truth: used by the sync cron, the partial index, and any +# code that needs to know whether a run is "done". +TERMINAL_STATUSES = ("completed", "failed", "terminated", "canceled", "timed_out") class TaskRunRequest(BaseModel): @@ -672,3 +678,31 @@ RunResponse = Annotated[Union[TaskRunResponse, WorkflowRunResponse], Field(discr class BlockRunResponse(WorkflowRunResponse): block_labels: list[str] = Field(description="A whitelist of block labels; only these blocks will execute") + + +class TaskRunListItem(BaseModel): + """Lightweight run-history item backed by the task_runs table.""" + + model_config = ConfigDict(from_attributes=True) + + task_run_id: str + run_id: str + task_run_type: str + status: str + title: str | None = None + started_at: datetime | None = None + finished_at: datetime | None = None + created_at: datetime + workflow_permanent_id: str | None = None + script_run: bool = False + searchable_text: str | None = Field(default=None, exclude=True) + + @field_validator("script_run", mode="before") + @classmethod + def coerce_script_run(cls, v: Any) -> bool: + """Intentionally lossy: collapse dict metadata / bool / None → bool for the list view. + + The full script execution metadata (dict) is available via the detail + endpoint's Run.script_run field. Do not rely on dict contents here. + """ + return bool(v) diff --git a/skyvern/services/task_v1_service.py b/skyvern/services/task_v1_service.py index 929b654ea..9befa77c4 100644 --- a/skyvern/services/task_v1_service.py +++ b/skyvern/services/task_v1_service.py @@ -14,7 +14,7 @@ from skyvern.forge.sdk.executor.factory import AsyncExecutorFactory from skyvern.forge.sdk.schemas.organizations import Organization from skyvern.forge.sdk.schemas.task_generations import TaskGeneration, TaskGenerationBase from skyvern.forge.sdk.schemas.tasks import Task, TaskRequest, TaskResponse, TaskStatus -from skyvern.schemas.runs import RunEngine, RunType +from skyvern.schemas.runs import RunEngine, RunStatus, RunType LOG = structlog.get_logger() @@ -101,6 +101,7 @@ async def run_task( title=task.title, url=task.url, url_hash=url_hash, + status=RunStatus.queued, ) if x_max_steps_override: LOG.info( diff --git a/skyvern/services/task_v2_service.py b/skyvern/services/task_v2_service.py index 3f275bc14..d55a114ae 100644 --- a/skyvern/services/task_v2_service.py +++ b/skyvern/services/task_v2_service.py @@ -40,7 +40,15 @@ from skyvern.forge.sdk.workflow.models.block import ( ) from skyvern.forge.sdk.workflow.models.parameter import PARAMETER_TYPE, ContextParameter from skyvern.forge.sdk.workflow.models.workflow import Workflow, WorkflowRequestBody, WorkflowRun, WorkflowRunStatus -from skyvern.schemas.runs import ProxyLocation, ProxyLocationInput, RunEngine, RunType, TaskRunRequest, TaskRunResponse +from skyvern.schemas.runs import ( + ProxyLocation, + ProxyLocationInput, + RunEngine, + RunStatus, + RunType, + TaskRunRequest, + TaskRunResponse, +) from skyvern.schemas.workflows import ( BLOCK_YAML_TYPES, PARAMETER_YAML_TYPES, @@ -334,6 +342,9 @@ async def initialize_task_v2( organization_id=organization.organization_id, run_id=task_v2.observer_cruise_id, title=new_workflow.title, + url=task_v2.url, + url_hash=generate_url_hash(task_v2.url) if task_v2.url else None, + status=RunStatus.queued, ) except Exception: LOG.warning("Failed to update task 2.0", exc_info=True) diff --git a/skyvern/services/workflow_service.py b/skyvern/services/workflow_service.py index 5ff20d21f..a1f41237b 100644 --- a/skyvern/services/workflow_service.py +++ b/skyvern/services/workflow_service.py @@ -60,6 +60,10 @@ async def prepare_workflow( organization_id=organization.organization_id, run_id=workflow_run.workflow_run_id, title=workflow.title, + status=RunStatus.queued, + workflow_permanent_id=workflow_id, + parent_workflow_run_id=parent_workflow_run_id, + debug_session_id=debug_session_id, ) if max_steps: diff --git a/tests/unit/embedded/conftest.py b/tests/unit/embedded/conftest.py index 17f3667ff..6e43a7ee1 100644 --- a/tests/unit/embedded/conftest.py +++ b/tests/unit/embedded/conftest.py @@ -9,50 +9,31 @@ DATABASE_STRING to SQLite, and subsequent scenario tests (which expect Postgres) fail with "no such table: organizations". """ -import copy +import os import pytest -from skyvern.config import settings +from skyvern.library.embedded_server_factory import _restore_settings, _snapshot_settings @pytest.fixture(autouse=True) -def _restore_settings(): - """Save and restore global settings around each embedded test. +def _restore_settings_fixture(): # type: ignore[no-untyped-def] # pytest fixture + """Save and restore global settings + forge app around each embedded test. - Embedded mode mutates the global settings singleton (DATABASE_STRING, - ADDITIONAL_MODULES, OTEL_ENABLED, etc.). Without restoration, subsequent - Postgres-based scenario tests in the same pytest process see SQLite - settings and fail. + Uses the shared snapshot/restore helpers from embedded_server_factory so + that the key list stays in sync with what _apply_sqlite_overrides mutates. """ - from skyvern.forge.sdk.settings_manager import SettingsManager + from skyvern.forge import app as forge_app_holder - keys_to_save = [ - "DATABASE_STRING", - "DATABASE_REPLICA_STRING", - "ADDITIONAL_MODULES", - "OTEL_ENABLED", - "LLM_KEY", - "BROWSER_LOGS_ENABLED", - ] - - original_values = {} - for key in keys_to_save: - val = getattr(settings, key, None) - original_values[key] = copy.deepcopy(val) if isinstance(val, (list, dict)) else val - - mgr_originals = {} - mgr_settings = SettingsManager.get_settings() - if mgr_settings is not settings: - for key in keys_to_save: - if hasattr(mgr_settings, key): - val = getattr(mgr_settings, key) - mgr_originals[key] = copy.deepcopy(val) if isinstance(val, (list, dict)) else val + snapshots = _snapshot_settings() + prev_app_inst = object.__getattribute__(forge_app_holder, "_inst") # type: ignore[arg-type] + prev_api_key = os.environ.get("SKYVERN_API_KEY") yield - for key, val in original_values.items(): - setattr(settings, key, val) - if mgr_settings is not settings: - for key, val in mgr_originals.items(): - setattr(mgr_settings, key, val) + _restore_settings(snapshots) + forge_app_holder.set_app(prev_app_inst) # type: ignore[attr-defined] + if prev_api_key is None: + os.environ.pop("SKYVERN_API_KEY", None) + else: + os.environ["SKYVERN_API_KEY"] = prev_api_key diff --git a/tests/unit/embedded/test_sqlite_server_mode.py b/tests/unit/embedded/test_sqlite_server_mode.py new file mode 100644 index 000000000..48c06074e --- /dev/null +++ b/tests/unit/embedded/test_sqlite_server_mode.py @@ -0,0 +1,430 @@ +"""Tests for SQLite-first server mode. + +Covers: +- _default_database_string() returns SQLite path under ~/.skyvern/ +- Settings.is_sqlite() detection +- SQLite bootstrap in api_app lifespan (tables, org, token, idempotency) +- Settings snapshot/restore on bootstrap failure +- ForgeApp restoration on bootstrap failure +""" + +from pathlib import Path +from unittest.mock import patch + +import httpx +import pytest +import pytest_asyncio + +from skyvern.config import _default_database_string + + +def test_default_database_string_is_sqlite() -> None: + """_default_database_string returns a SQLite URL pointing at ~/.skyvern/data.db.""" + result = _default_database_string() + assert result.startswith("sqlite+aiosqlite:///") + assert ".skyvern/data.db" in result + + +def test_default_database_string_is_pure(tmp_path: Path) -> None: + """_default_database_string is a pure string computation — no side effects.""" + fake_home = tmp_path / "fakehome" + with patch("skyvern.config.Path.home", return_value=fake_home): + result = _default_database_string() + assert not (fake_home / ".skyvern").exists(), "factory must not create directories" + assert ".skyvern/data.db" in result + + +def test_ensure_sqlite_dir_creates_directory(tmp_path: Path) -> None: + """_ensure_sqlite_dir creates the parent directory for file-backed SQLite.""" + from skyvern.config import _ensure_sqlite_dir + + db_path = tmp_path / "subdir" / "data.db" + _ensure_sqlite_dir(f"sqlite+aiosqlite:///{db_path}") + assert (tmp_path / "subdir").is_dir() + + +def test_ensure_sqlite_dir_noop_for_memory() -> None: + """_ensure_sqlite_dir is a no-op for in-memory SQLite.""" + from skyvern.config import _ensure_sqlite_dir + + _ensure_sqlite_dir("sqlite+aiosqlite:///:memory:") # should not raise + + +def test_ensure_sqlite_dir_noop_for_postgres() -> None: + """_ensure_sqlite_dir is a no-op for Postgres URLs.""" + from skyvern.config import _ensure_sqlite_dir + + _ensure_sqlite_dir("postgresql+psycopg://localhost/test") # should not raise + + +def test_is_sqlite_true_for_sqlite_string() -> None: + """is_sqlite() returns True when DATABASE_STRING starts with 'sqlite'.""" + from skyvern.config import Settings + + s = Settings(DATABASE_STRING="sqlite+aiosqlite:///test.db") + assert s.is_sqlite() is True + + +def test_is_sqlite_false_for_postgres_string() -> None: + """is_sqlite() returns False for PostgreSQL strings.""" + from skyvern.config import Settings + + s = Settings(DATABASE_STRING="postgresql+psycopg://skyvern@localhost/skyvern") + assert s.is_sqlite() is False + + +@pytest_asyncio.fixture +async def sqlite_bootstrap_db(): # type: ignore[no-untyped-def] # pytest fixture + """Swap in a disposable SQLite AgentDB for bootstrap tests.""" + from skyvern.forge import app as forge_app + from skyvern.forge.sdk.db.agent_db import AgentDB + + db = AgentDB("sqlite+aiosqlite:///:memory:") + original_db = forge_app.DATABASE + forge_app.DATABASE = db # type: ignore[assignment] + try: + yield db + finally: + forge_app.DATABASE = original_db # type: ignore[assignment] + await db.engine.dispose() + + +@pytest.fixture +def patched_env_writes(): # type: ignore[no-untyped-def] # pytest fixture + """Mock env-file writes so bootstrap tests do not touch the repo .env.""" + with patch("skyvern.forge.sdk.services.local_org_auth_token_service._write_env") as write_env: + yield write_env + + +@pytest.mark.asyncio +async def test_sqlite_bootstrap_creates_tables_and_org(sqlite_bootstrap_db, patched_env_writes) -> None: + """_bootstrap_sqlite creates tables, org, and API key in a SQLite DB.""" + from skyvern.forge.api_app import _bootstrap_sqlite + from skyvern.forge.sdk.db.enums import OrganizationAuthTokenType + from skyvern.forge.sdk.db.models import Base + from skyvern.forge.sdk.services.local_org_auth_token_service import SKYVERN_LOCAL_DOMAIN + + async with sqlite_bootstrap_db.engine.begin() as conn: + await conn.run_sync(Base.metadata.create_all) + + await _bootstrap_sqlite() + + org = await sqlite_bootstrap_db.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + assert org is not None + assert org.organization_name == "Skyvern-local" + token = await sqlite_bootstrap_db.get_valid_org_auth_token(org.organization_id, OrganizationAuthTokenType.api) + assert token is not None + + +@pytest.mark.asyncio +async def test_sqlite_bootstrap_is_idempotent(sqlite_bootstrap_db, patched_env_writes) -> None: + """Calling _bootstrap_sqlite twice does not create duplicate orgs.""" + from skyvern.forge.api_app import _bootstrap_sqlite + from skyvern.forge.sdk.db.enums import OrganizationAuthTokenType + from skyvern.forge.sdk.db.models import Base + from skyvern.forge.sdk.services.local_org_auth_token_service import SKYVERN_LOCAL_DOMAIN + + async with sqlite_bootstrap_db.engine.begin() as conn: + await conn.run_sync(Base.metadata.create_all) + + await _bootstrap_sqlite() + org1 = await sqlite_bootstrap_db.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + + # Second call should detect existing org and skip + await _bootstrap_sqlite() + org2 = await sqlite_bootstrap_db.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + + assert org1 is not None + assert org2 is not None + assert org1.organization_id == org2.organization_id + token = await sqlite_bootstrap_db.get_valid_org_auth_token(org1.organization_id, OrganizationAuthTokenType.api) + assert token is not None + + +@pytest.mark.asyncio +async def test_sqlite_bootstrap_from_empty_db(sqlite_bootstrap_db, patched_env_writes) -> None: + """_bootstrap_sqlite creates tables AND org from a completely empty DB. + + Unlike test_sqlite_bootstrap_creates_tables_and_org which pre-creates + tables, this test starts from scratch to cover the full first-start path. + """ + from skyvern.forge.api_app import _bootstrap_sqlite + from skyvern.forge.sdk.services.local_org_auth_token_service import SKYVERN_LOCAL_DOMAIN + + # NO create_all — bootstrap must handle it + + await _bootstrap_sqlite() + + org = await sqlite_bootstrap_db.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + assert org is not None + assert org.organization_name == "Skyvern-local" + + +@pytest.mark.asyncio +async def test_sqlite_bootstrap_syncs_existing_env_api_key( + monkeypatch: pytest.MonkeyPatch, + tmp_path: Path, + sqlite_bootstrap_db, + patched_env_writes, +) -> None: + """An existing SKYVERN_API_KEY must become a valid token in a fresh SQLite DB.""" + from skyvern.forge.api_app import _bootstrap_sqlite + from skyvern.forge.sdk.core.security import create_access_token + from skyvern.forge.sdk.db.enums import OrganizationAuthTokenType + from skyvern.forge.sdk.services.local_org_auth_token_service import SKYVERN_LOCAL_DOMAIN + from skyvern.forge.sdk.services.org_auth_service import resolve_org_from_api_key + + monkeypatch.chdir(tmp_path) + expected_org_id = "o_existing_local" + existing_api_key = create_access_token(expected_org_id) + monkeypatch.setenv("SKYVERN_API_KEY", existing_api_key) + # Bootstrap reads settings.SKYVERN_API_KEY (the pydantic singleton), not os.environ directly + monkeypatch.setattr("skyvern.config.settings.SKYVERN_API_KEY", existing_api_key) + + await _bootstrap_sqlite() + + org = await sqlite_bootstrap_db.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + assert org is not None + assert org.organization_id == expected_org_id + token = await sqlite_bootstrap_db.get_valid_org_auth_token(org.organization_id, OrganizationAuthTokenType.api) + assert token is not None + assert token.token == existing_api_key + validation = await resolve_org_from_api_key(existing_api_key, sqlite_bootstrap_db) + assert validation.organization.organization_id == expected_org_id + patched_env_writes.assert_not_called() + + +@pytest.mark.asyncio +async def test_sqlite_bootstrap_repairs_existing_org_without_token( + monkeypatch: pytest.MonkeyPatch, + tmp_path: Path, + sqlite_bootstrap_db, + patched_env_writes, +) -> None: + """Bootstrap should self-heal an existing local org that has no API token.""" + from skyvern.forge.api_app import _bootstrap_sqlite + from skyvern.forge.sdk.core.security import create_access_token + from skyvern.forge.sdk.db.enums import OrganizationAuthTokenType + from skyvern.forge.sdk.db.models import Base + from skyvern.forge.sdk.services.local_org_auth_token_service import ( + SKYVERN_LOCAL_DOMAIN, + ensure_local_org_with_id, + ) + from skyvern.forge.sdk.services.org_auth_service import resolve_org_from_api_key + + monkeypatch.chdir(tmp_path) + expected_org_id = "o_existing_local" + existing_api_key = create_access_token(expected_org_id) + monkeypatch.setenv("SKYVERN_API_KEY", existing_api_key) + monkeypatch.setattr("skyvern.config.settings.SKYVERN_API_KEY", existing_api_key) + + async with sqlite_bootstrap_db.engine.begin() as conn: + await conn.run_sync(Base.metadata.create_all) + + org = await ensure_local_org_with_id(expected_org_id) + assert org.organization_id + assert ( + await sqlite_bootstrap_db.get_valid_org_auth_token(org.organization_id, OrganizationAuthTokenType.api) is None + ) + + await _bootstrap_sqlite() + + repaired_org = await sqlite_bootstrap_db.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + assert repaired_org is not None + token = await sqlite_bootstrap_db.get_valid_org_auth_token( + repaired_org.organization_id, OrganizationAuthTokenType.api + ) + assert token is not None + assert token.token == existing_api_key + validation = await resolve_org_from_api_key(existing_api_key, sqlite_bootstrap_db) + assert validation.organization.organization_id == expected_org_id + patched_env_writes.assert_not_called() + + +@pytest.mark.asyncio +async def test_sqlite_bootstrap_regenerates_invalid_existing_env_api_key( + monkeypatch: pytest.MonkeyPatch, + tmp_path: Path, + sqlite_bootstrap_db, + patched_env_writes, +) -> None: + """Bootstrap must replace an unusable env key with a valid local JWT.""" + from skyvern.forge.api_app import _bootstrap_sqlite + from skyvern.forge.sdk.db.enums import OrganizationAuthTokenType + from skyvern.forge.sdk.services.local_org_auth_token_service import SKYVERN_LOCAL_DOMAIN + from skyvern.forge.sdk.services.org_auth_service import resolve_org_from_api_key + + monkeypatch.chdir(tmp_path) + monkeypatch.setenv("SKYVERN_API_KEY", "existing-test-key") + + await _bootstrap_sqlite() + + org = await sqlite_bootstrap_db.get_organization_by_domain(SKYVERN_LOCAL_DOMAIN) + assert org is not None + token = await sqlite_bootstrap_db.get_valid_org_auth_token(org.organization_id, OrganizationAuthTokenType.api) + assert token is not None + assert token.token != "existing-test-key" + validation = await resolve_org_from_api_key(token.token, sqlite_bootstrap_db) + assert validation.organization.organization_id == org.organization_id + + +@pytest.mark.asyncio +async def test_local_allows_env_only_persistent_mode(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> None: + """Skyvern.local() should honor env-only persistent mode without requiring ./.env.""" + from skyvern import Skyvern + + monkeypatch.chdir(tmp_path) + monkeypatch.setenv("DATABASE_STRING", "postgresql+psycopg://skyvern@localhost/skyvern") + monkeypatch.setenv("SKYVERN_API_KEY", "dummy-key") + + embedded_client = httpx.AsyncClient() + with patch( + "skyvern.library.embedded_server_factory.create_embedded_server", return_value=embedded_client + ) as factory: + skyvern = Skyvern.local() + + try: + assert factory.call_args.kwargs["use_in_memory_db"] is False + assert getattr(skyvern, "_embedded_client") is embedded_client + finally: + await skyvern.aclose() + + +@pytest.mark.asyncio +async def test_local_persistent_mode_accepts_settings_without_dotenv( + monkeypatch: pytest.MonkeyPatch, tmp_path: Path +) -> None: + """Persistent mode should accept DATABASE_STRING and SKYVERN_API_KEY via settings overrides.""" + from skyvern import Skyvern + + monkeypatch.chdir(tmp_path) + + embedded_client = httpx.AsyncClient() + overrides = { + "DATABASE_STRING": "postgresql+psycopg://skyvern@localhost/skyvern", + "SKYVERN_API_KEY": "dummy-key", + } + with patch( + "skyvern.library.embedded_server_factory.create_embedded_server", return_value=embedded_client + ) as factory: + skyvern = Skyvern.local(use_in_memory_db=False, settings=overrides) + + try: + assert factory.call_args.kwargs["use_in_memory_db"] is False + assert factory.call_args.kwargs["settings_overrides"] == overrides + assert getattr(skyvern, "_embedded_client") is embedded_client + finally: + await skyvern.aclose() + + +@pytest.mark.asyncio +async def test_create_embedded_server_uses_settings_api_key_in_persistent_mode() -> None: + """Persistent embedded bootstrap should read SKYVERN_API_KEY from settings overrides.""" + from skyvern.library.embedded_server_factory import create_embedded_server + + seen_headers: dict[bytes, bytes] = {} + + async def fake_app(scope, receive, send): # type: ignore[no-untyped-def] + nonlocal seen_headers + seen_headers = dict(scope["headers"]) + await send( + { + "type": "http.response.start", + "status": 200, + "headers": [(b"content-type", b"application/json")], + } + ) + await send({"type": "http.response.body", "body": b"[]"}) + + client = create_embedded_server( + settings_overrides={"SKYVERN_API_KEY": "dummy-key"}, + use_in_memory_db=False, + ) + try: + with patch("skyvern.library.embedded_server_factory.create_api_app", return_value=fake_app): + response = await client.get("/") + finally: + await client.aclose() + + assert response.status_code == 200 + assert seen_headers[b"x-api-key"] == b"dummy-key" + + +def test_settings_snapshot_restore_roundtrip() -> None: + """Snapshot captures values and restore puts them back after mutation.""" + from skyvern.config import settings + from skyvern.library.embedded_server_factory import _restore_settings, _snapshot_settings + + original_db = settings.DATABASE_STRING + original_modules = settings.ADDITIONAL_MODULES[:] + + snapshots = _snapshot_settings() + + # Mutate settings + settings.DATABASE_STRING = "sqlite+aiosqlite:///:memory:" + settings.ADDITIONAL_MODULES = [] + + # Restore + _restore_settings(snapshots) + + assert settings.DATABASE_STRING == original_db + assert settings.ADDITIONAL_MODULES == original_modules + + +def test_settings_snapshot_keys_cover_embedded_mutations() -> None: + """Snapshot coverage should include both SQLite overrides and bootstrap-time mutations.""" + from skyvern.library.embedded_server_factory import ( + _BOOTSTRAP_RUNTIME_SETTINGS, + _SETTINGS_SNAPSHOT_KEYS, + _SQLITE_OVERRIDE_VALUES, + ) + + assert frozenset(_SQLITE_OVERRIDE_VALUES).issubset(_SETTINGS_SNAPSHOT_KEYS) + assert _BOOTSTRAP_RUNTIME_SETTINGS.issubset(_SETTINGS_SNAPSHOT_KEYS) + + +@pytest.mark.asyncio +async def test_bootstrap_failure_restores_settings() -> None: + """If bootstrap fails, settings must be restored to pre-bootstrap values.""" + from skyvern.config import settings + + original_db = settings.DATABASE_STRING + + from skyvern import Skyvern + + # Create a client with a bad setting that will cause validation error + skyvern = Skyvern.local( + use_in_memory_db=True, + settings={"OTEL_ENABLED": True}, # Blocked setting + ) + + with pytest.raises(ValueError, match="Cannot override"): + await skyvern.get_workflows() + + await skyvern.aclose() + + # Settings should be restored to original values + assert settings.DATABASE_STRING == original_db + + +@pytest.mark.asyncio +async def test_bootstrap_failure_restores_forge_app() -> None: + """If bootstrap fails, the forge app instance must be restored.""" + from skyvern.forge import app as forge_app_holder + + prev_inst = object.__getattribute__(forge_app_holder, "_inst") + + from skyvern import Skyvern + + skyvern = Skyvern.local( + use_in_memory_db=True, + settings={"OTEL_ENABLED": True}, # Blocked setting + ) + + with pytest.raises(ValueError, match="Cannot override"): + await skyvern.get_workflows() + + await skyvern.aclose() + + current_inst = object.__getattribute__(forge_app_holder, "_inst") + assert current_inst is prev_inst diff --git a/tests/unit/test_task_run_status_sync.py b/tests/unit/test_task_run_status_sync.py new file mode 100644 index 000000000..5c3353f9f --- /dev/null +++ b/tests/unit/test_task_run_status_sync.py @@ -0,0 +1,64 @@ +"""Tests for task_run status write-through sync.""" + +from unittest.mock import AsyncMock, MagicMock + +import pytest + +from skyvern.forge.sdk.db.repositories.tasks import TasksRepository + + +@pytest.fixture +def mock_session(): + session = AsyncMock() + session.__aenter__ = AsyncMock(return_value=session) + session.__aexit__ = AsyncMock(return_value=False) + return session + + +@pytest.fixture +def tasks_repo(mock_session): + repo = TasksRepository.__new__(TasksRepository) + repo.Session = MagicMock(return_value=mock_session) + repo.debug_enabled = False + repo._is_retryable_error_fn = None + return repo + + +@pytest.mark.asyncio +async def test_sync_task_run_status_updates_matching_row(tasks_repo, mock_session): + """sync_task_run_status should UPDATE task_runs where run_id matches.""" + await tasks_repo.sync_task_run_status( + organization_id="org_1", + run_id="wr_123", + status="failed", + ) + mock_session.execute.assert_called_once() + call_args = mock_session.execute.call_args + # The SQL should be an UPDATE on task_runs + sql_text = str(call_args[0][0]) + assert "task_runs" in sql_text + assert "status" in sql_text + + +@pytest.mark.asyncio +async def test_sync_task_run_status_no_raise_on_error(tasks_repo, mock_session): + """sync_task_run_status should swallow exceptions (best-effort).""" + mock_session.execute.side_effect = Exception("DB error") + # Should NOT raise + await tasks_repo.sync_task_run_status( + organization_id="org_1", + run_id="nonexistent", + status="failed", + ) + + +def test_terminal_statuses_match_run_status(): + """Guard: TERMINAL_STATUSES and RunStatus.is_final() must agree. + + If this fails, a new terminal status was added to one but not the other. + Update TERMINAL_STATUSES in skyvern/schemas/runs.py (the single source of truth). + """ + from skyvern.forge.sdk.schemas.runs import TERMINAL_STATUSES + from skyvern.schemas.runs import RunStatus + + assert set(TERMINAL_STATUSES) == {s.value for s in RunStatus if s.is_final()} diff --git a/tests/unit/test_task_v2_task_run_creation.py b/tests/unit/test_task_v2_task_run_creation.py new file mode 100644 index 000000000..b2ce08c3d --- /dev/null +++ b/tests/unit/test_task_v2_task_run_creation.py @@ -0,0 +1,51 @@ +from types import SimpleNamespace + +import pytest + +from skyvern.forge import app +from skyvern.forge.sdk.core.hashing import generate_url_hash +from skyvern.schemas.runs import RunStatus, RunType +from skyvern.services.task_v2_service import DEFAULT_WORKFLOW_TITLE, initialize_task_v2 + + +@pytest.mark.asyncio +async def test_initialize_task_v2_populates_task_run_url_when_user_url_is_known() -> None: + organization = SimpleNamespace(organization_id="org_123") + user_url = "https://example.com" + + app.DATABASE.create_task_v2.return_value = SimpleNamespace( + observer_cruise_id="tsk_123", + workflow_run_id=None, + url=user_url, + ) + app.WORKFLOW_SERVICE.create_empty_workflow.return_value = SimpleNamespace( + workflow_id="wf_123", + workflow_permanent_id="wpid_123", + title=DEFAULT_WORKFLOW_TITLE, + ) + app.WORKFLOW_SERVICE.setup_workflow_run.return_value = SimpleNamespace(workflow_run_id="wr_123") + app.DATABASE.update_task_v2.return_value = SimpleNamespace( + observer_cruise_id="tsk_123", + workflow_run_id="wr_123", + workflow_id="wf_123", + workflow_permanent_id="wpid_123", + url=user_url, + ) + app.DATABASE.create_task_run.return_value = SimpleNamespace(run_id="tsk_123") + + await initialize_task_v2( + organization=organization, + user_prompt="Open the page", + user_url=user_url, + create_task_run=True, + ) + + app.DATABASE.create_task_run.assert_awaited_once_with( + task_run_type=RunType.task_v2, + organization_id="org_123", + run_id="tsk_123", + title=DEFAULT_WORKFLOW_TITLE, + url=user_url, + url_hash=generate_url_hash(user_url), + status=RunStatus.queued, + ) diff --git a/uv.lock b/uv.lock index 7c72fed05..ec9f33df7 100644 --- a/uv.lock +++ b/uv.lock @@ -7,6 +7,10 @@ resolution-markers = [ "python_full_version < '3.12'", ] +[options] +exclude-newer = "2026-03-24T13:53:47.750565Z" +exclude-newer-span = "P7D" + [manifest] constraints = [ { name = "authlib", specifier = ">=1.6.9" }, @@ -654,16 +658,16 @@ wheels = [ [[package]] name = "build" -version = "1.4.2" +version = "1.4.0" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "colorama", marker = "os_name == 'nt'" }, { name = "packaging" }, { name = "pyproject-hooks" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/6c/1d/ab15c8ac57f4ee8778d7633bc6685f808ab414437b8644f555389cdc875e/build-1.4.2.tar.gz", hash = "sha256:35b14e1ee329c186d3f08466003521ed7685ec15ecffc07e68d706090bf161d1", size = 83433, upload-time = "2026-03-25T14:20:27.659Z" } +sdist = { url = "https://files.pythonhosted.org/packages/42/18/94eaffda7b329535d91f00fe605ab1f1e5cd68b2074d03f255c7d250687d/build-1.4.0.tar.gz", hash = "sha256:f1b91b925aa322be454f8330c6fb48b465da993d1e7e7e6fa35027ec49f3c936", size = 50054, upload-time = "2026-01-08T16:41:47.696Z" } wheels = [ - { url = "https://files.pythonhosted.org/packages/4a/57/3b7d4dd193ade4641c865bc2b93aeeb71162e81fc348b8dad020215601ed/build-1.4.2-py3-none-any.whl", hash = "sha256:7a4d8651ea877cb2a89458b1b198f2e69f536c95e89129dbf5d448045d60db88", size = 24643, upload-time = "2026-03-25T14:20:26.568Z" }, + { url = "https://files.pythonhosted.org/packages/c5/0d/84a4380f930db0010168e0aa7b7a8fed9ba1835a8fbb1472bc6d0201d529/build-1.4.0-py3-none-any.whl", hash = "sha256:6a07c1b8eb6f2b311b96fcbdbce5dab5fe637ffda0fd83c9cac622e927501596", size = 24141, upload-time = "2026-01-08T16:41:46.453Z" }, ] [[package]] @@ -764,7 +768,7 @@ wheels = [ [[package]] name = "cfn-lint" -version = "1.47.1" +version = "1.47.0" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "aws-sam-translator" }, @@ -775,9 +779,9 @@ dependencies = [ { name = "sympy" }, { name = "typing-extensions" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/10/34/e66811016e7709cab78b0cf896437b922d7537986ac727344663b6cc2044/cfn_lint-1.47.1.tar.gz", hash = "sha256:b2eedbcee3aa104602f79933e3ad74c01f0fa1e226b70327118926fd78d8d3f1", size = 3672271, upload-time = "2026-03-24T15:59:34.526Z" } +sdist = { url = "https://files.pythonhosted.org/packages/61/f2/67074ab011d47e973b57e1efb6adf798fbaab3873b9a3edcf6f8b9c33639/cfn_lint-1.47.0.tar.gz", hash = "sha256:29a9377febd6c8fa30903d437a010bcf042426b220d009101c6cf8523819291f", size = 3673549, upload-time = "2026-03-17T17:14:45.95Z" } wheels = [ - { url = "https://files.pythonhosted.org/packages/a5/88/19802ef0e1ef6259c4bc4b58226c0e7ff8b7ae93806ca32354c007e3480a/cfn_lint-1.47.1-py3-none-any.whl", hash = "sha256:3a4b5dba0fd03c24f2bc0e112a88ad90fa29014971e881b8f1e297d22f398a97", size = 5299292, upload-time = "2026-03-24T15:59:31.86Z" }, + { url = "https://files.pythonhosted.org/packages/16/2e/c552923e89004a268f94520de836525523132dd76ec52d3476a076d9dda2/cfn_lint-1.47.0-py3-none-any.whl", hash = "sha256:1148f37f2733cdb361897746eda903b5414ad5dd7e5df9ea634ac9f6c67975c9", size = 5281656, upload-time = "2026-03-17T17:14:43.607Z" }, ] [[package]] @@ -1412,7 +1416,7 @@ wheels = [ [[package]] name = "google-cloud-aiplatform" -version = "1.143.0" +version = "1.142.0" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "docstring-parser" }, @@ -1428,9 +1432,9 @@ dependencies = [ { name = "pydantic" }, { name = "typing-extensions" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/a7/08/939fb05870fdf155410a927e22f5b053d49f18e215618e102fba1d8bb147/google_cloud_aiplatform-1.143.0.tar.gz", hash = "sha256:1f0124a89795a6b473deb28724dd37d95334205df3a9c9c48d0b8d7a3d5d5cc4", size = 10215389, upload-time = "2026-03-25T18:30:15.444Z" } +sdist = { url = "https://files.pythonhosted.org/packages/41/0d/3063a0512d60cf18854a279e00ccb796429545464345ef821cf77cb93d05/google_cloud_aiplatform-1.142.0.tar.gz", hash = "sha256:87b49e002703dc14885093e9b264587db84222bef5f70f5a442d03f41beecdd1", size = 10207993, upload-time = "2026-03-20T22:49:13.797Z" } wheels = [ - { url = "https://files.pythonhosted.org/packages/90/14/16323e604e79dc63b528268f97a841c2c29dd8eb16395de6bf530c1a5ebe/google_cloud_aiplatform-1.143.0-py2.py3-none-any.whl", hash = "sha256:78df97d044859f743a9cc48b89a260d33579b0d548b1589bb3ae9f4c2afc0c5a", size = 8392705, upload-time = "2026-03-25T18:30:11.496Z" }, + { url = "https://files.pythonhosted.org/packages/59/8b/f29646d3fa940f0e38cfcc12137f4851856b50d7486a3c05103ebc78d82d/google_cloud_aiplatform-1.142.0-py2.py3-none-any.whl", hash = "sha256:17c91db9b613cbbafb2c36335b123686aeb2b4b8448be5134b565ae07165a39a", size = 8388991, upload-time = "2026-03-20T22:49:10.334Z" }, ] [[package]] @@ -1900,7 +1904,7 @@ wheels = [ [[package]] name = "huggingface-hub" -version = "1.8.0" +version = "1.7.2" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "filelock" }, @@ -1913,9 +1917,9 @@ dependencies = [ { name = "typer" }, { name = "typing-extensions" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/8e/2a/a847fd02261cd051da218baf99f90ee7c7040c109a01833db4f838f25256/huggingface_hub-1.8.0.tar.gz", hash = "sha256:c5627b2fd521e00caf8eff4ac965ba988ea75167fad7ee72e17f9b7183ec63f3", size = 735839, upload-time = "2026-03-25T16:01:28.152Z" } +sdist = { url = "https://files.pythonhosted.org/packages/19/15/eafc1c57bf0f8afffb243dcd4c0cceb785e956acc17bba4d9bf2ae21fc9c/huggingface_hub-1.7.2.tar.gz", hash = "sha256:7f7e294e9bbb822e025bdb2ada025fa4344d978175a7f78e824d86e35f7ab43b", size = 724684, upload-time = "2026-03-20T10:36:08.767Z" } wheels = [ - { url = "https://files.pythonhosted.org/packages/a9/ae/8a3a16ea4d202cb641b51d2681bdd3d482c1c592d7570b3fa264730829ce/huggingface_hub-1.8.0-py3-none-any.whl", hash = "sha256:d3eb5047bd4e33c987429de6020d4810d38a5bef95b3b40df9b17346b7f353f2", size = 625208, upload-time = "2026-03-25T16:01:26.603Z" }, + { url = "https://files.pythonhosted.org/packages/08/de/3ad061a05f74728927ded48c90b73521b9a9328c85d841bdefb30e01fb85/huggingface_hub-1.7.2-py3-none-any.whl", hash = "sha256:288f33a0a17b2a73a1359e2a5fd28d1becb2c121748c6173ab8643fb342c850e", size = 618036, upload-time = "2026-03-20T10:36:06.824Z" }, ] [[package]] @@ -3192,26 +3196,26 @@ wheels = [ [[package]] name = "nh3" -version = "0.3.4" +version = "0.3.3" source = { registry = "https://pypi.org/simple" } -sdist = { url = "https://files.pythonhosted.org/packages/4e/86/f8d3a7c9bd1bbaa181f6312c757e0b74d25f71ecf84ea3c0dc5e0f01840d/nh3-0.3.4.tar.gz", hash = "sha256:96709a379997c1b28c8974146ca660b0dcd3794f4f6d50c1ea549bab39ac6ade", size = 19520, upload-time = "2026-03-25T10:57:30.789Z" } +sdist = { url = "https://files.pythonhosted.org/packages/cc/37/ab55eb2b05e334ff9a1ad52c556ace1f9c20a3f63613a165d384d5387657/nh3-0.3.3.tar.gz", hash = "sha256:185ed41b88c910b9ca8edc89ca3b4be688a12cb9de129d84befa2f74a0039fee", size = 18968, upload-time = "2026-02-14T09:35:15.664Z" } wheels = [ - { url = "https://files.pythonhosted.org/packages/4a/57/a97955bc95960cfb1f0517043d60a121f4ba93fde252d4d9ffd3c2a9eead/nh3-0.3.4-cp38-abi3-macosx_10_12_x86_64.macosx_11_0_arm64.macosx_10_12_universal2.whl", hash = "sha256:d8bebcb20ab4b91858385cd98fe58046ec4a624275b45ef9b976475604f45b49", size = 1439519, upload-time = "2026-03-25T10:57:12.019Z" }, - { url = "https://files.pythonhosted.org/packages/2b/60/c9a33361da8cde7c7760f091cd10467bc470634e4eea31c8bb70935b00a4/nh3-0.3.4-cp38-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:0d825722a1e8cbc87d7ca1e47ffb1d2a6cf343ad4c1b8465becf7cadcabcdfd0", size = 833798, upload-time = "2026-03-25T10:57:13.264Z" }, - { url = "https://files.pythonhosted.org/packages/6b/19/9487790780b8c94eacca37866c1270b747a4af8e244d43b3b550fddbbf62/nh3-0.3.4-cp38-abi3-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:4aa8b43e68c26b68069a3b6cef09de166d1d7fa140cf8d77e409a46cbf742e44", size = 820414, upload-time = "2026-03-25T10:57:14.236Z" }, - { url = "https://files.pythonhosted.org/packages/6b/b4/c6a340dd321d20b1e4a663307032741da045685c87403926c43656f6f5ec/nh3-0.3.4-cp38-abi3-manylinux_2_17_ppc64.manylinux2014_ppc64.whl", hash = "sha256:f5f214618ad5eff4f2a6b13a8d4da4d9e7f37c569d90a13fb9f0caaf7d04fe21", size = 1061531, upload-time = "2026-03-25T10:57:15.384Z" }, - { url = "https://files.pythonhosted.org/packages/c4/49/f6b4b474e0032e4bcbb7174b44e4cf6915670e09c62421deb06ccfcb88b8/nh3-0.3.4-cp38-abi3-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:3390e4333883673a684ce16c1716b481e91782d6f56dec5c85fed9feedb23382", size = 1021889, upload-time = "2026-03-25T10:57:16.454Z" }, - { url = "https://files.pythonhosted.org/packages/43/da/e52a6941746d1f974752af3fc8591f1dbcdcf7fd8c726c7d99f444ba820e/nh3-0.3.4-cp38-abi3-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:18a2e44ccb29cbb45071b8f3f2dab9ebfb41a6516f328f91f1f1fd18196239a4", size = 912965, upload-time = "2026-03-25T10:57:17.624Z" }, - { url = "https://files.pythonhosted.org/packages/d6/b7/ec1cbc6b297a808c513f59f501656389623fc09ad6a58c640851289c7854/nh3-0.3.4-cp38-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0961a27dc2057c38d0364cb05880e1997ae1c80220cbc847db63213720b8f304", size = 804975, upload-time = "2026-03-25T10:57:18.994Z" }, - { url = "https://files.pythonhosted.org/packages/a9/56/b1275aa2c6510191eed76178da4626b0900402439cb9f27d6b9bf7c6d5e9/nh3-0.3.4-cp38-abi3-manylinux_2_31_riscv64.whl", hash = "sha256:9337517edb7c10228252cce2898e20fb3d77e32ffaccbb3c66897927d74215a0", size = 833400, upload-time = "2026-03-25T10:57:20.086Z" }, - { url = "https://files.pythonhosted.org/packages/7c/a5/5d574ffa3c6e49a5364d1b25ebad165501c055340056671493beb467a15e/nh3-0.3.4-cp38-abi3-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:d866701affe67a5171b916b5c076e767a74c6a9efb7fb2006eb8d3c5f9a293d5", size = 854277, upload-time = "2026-03-25T10:57:21.433Z" }, - { url = "https://files.pythonhosted.org/packages/79/36/8aeb2ab21517cefa212db109e41024e02650716cb42bf293d0a88437a92d/nh3-0.3.4-cp38-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:47d749d99ae005ab19517224140b280dd56e77b33afb82f9b600e106d0458003", size = 1022021, upload-time = "2026-03-25T10:57:22.433Z" }, - { url = "https://files.pythonhosted.org/packages/9c/95/9fd860997685e64abe2d5a995ca2eb5004c0fb6d6585429612a7871548b9/nh3-0.3.4-cp38-abi3-musllinux_1_2_armv7l.whl", hash = "sha256:f987cb56458323405e8e5ea827e1befcf141ffa0c0ac797d6d02e6b646056d9a", size = 1103526, upload-time = "2026-03-25T10:57:23.487Z" }, - { url = "https://files.pythonhosted.org/packages/7d/0d/df545070614c1007f0109bb004230226c9000e7857c9785583ec25cda9d7/nh3-0.3.4-cp38-abi3-musllinux_1_2_i686.whl", hash = "sha256:883d5a6d6ee8078c4afc8e96e022fe579c4c265775ff6ee21e39b8c542cabab3", size = 1068050, upload-time = "2026-03-25T10:57:24.624Z" }, - { url = "https://files.pythonhosted.org/packages/94/d5/17b016df52df052f714c53be71df26a1943551d9931e9383b92c998b88f8/nh3-0.3.4-cp38-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:75643c22f5092d8e209f766ee8108c400bc1e44760fc94d2d638eb138d18f853", size = 1046037, upload-time = "2026-03-25T10:57:25.799Z" }, - { url = "https://files.pythonhosted.org/packages/51/39/49f737907e6ab2b4ca71855d3bd63dd7958862e9c8b94fb4e5b18ccf6988/nh3-0.3.4-cp38-abi3-win32.whl", hash = "sha256:72e4e9ca1c4bd41b4a28b0190edc2e21e3f71496acd36a0162858e1a28db3d7e", size = 609542, upload-time = "2026-03-25T10:57:27.112Z" }, - { url = "https://files.pythonhosted.org/packages/73/4f/af8e9071d7464575a7316831938237ffc9d92d27f163dbdd964b1309cd9b/nh3-0.3.4-cp38-abi3-win_amd64.whl", hash = "sha256:c10b1f0c741e257a5cb2978d6bac86e7c784ab20572724b20c6402c2e24bce75", size = 624244, upload-time = "2026-03-25T10:57:28.302Z" }, - { url = "https://files.pythonhosted.org/packages/44/0c/37695d6b0168f6714b5c492331636a9e6123d6ec22d25876c68d06eab1b8/nh3-0.3.4-cp38-abi3-win_arm64.whl", hash = "sha256:43ad4eedee7e049b9069bc015b7b095d320ed6d167ecec111f877de1540656e9", size = 616649, upload-time = "2026-03-25T10:57:29.623Z" }, + { url = "https://files.pythonhosted.org/packages/13/3e/aef8cf8e0419b530c95e96ae93a5078e9b36c1e6613eeb1df03a80d5194e/nh3-0.3.3-cp38-abi3-macosx_10_12_x86_64.macosx_11_0_arm64.macosx_10_12_universal2.whl", hash = "sha256:e8ee96156f7dfc6e30ecda650e480c5ae0a7d38f0c6fafc3c1c655e2500421d9", size = 1448640, upload-time = "2026-02-14T09:34:49.316Z" }, + { url = "https://files.pythonhosted.org/packages/ca/43/d2011a4f6c0272cb122eeff40062ee06bb2b6e57eabc3a5e057df0d582df/nh3-0.3.3-cp38-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:45fe0d6a607264910daec30360c8a3b5b1500fd832d21b2da608256287bcb92d", size = 839405, upload-time = "2026-02-14T09:34:50.779Z" }, + { url = "https://files.pythonhosted.org/packages/f8/f3/965048510c1caf2a34ed04411a46a04a06eb05563cd06f1aa57b71eb2bc8/nh3-0.3.3-cp38-abi3-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:5bc1d4b30ba1ba896669d944b6003630592665974bd11a3dc2f661bde92798a7", size = 825849, upload-time = "2026-02-14T09:34:52.622Z" }, + { url = "https://files.pythonhosted.org/packages/78/99/b4bbc6ad16329d8db2c2c320423f00b549ca3b129c2b2f9136be2606dbb0/nh3-0.3.3-cp38-abi3-manylinux_2_17_ppc64.manylinux2014_ppc64.whl", hash = "sha256:f433a2dd66545aad4a720ad1b2150edcdca75bfff6f4e6f378ade1ec138d5e77", size = 1068303, upload-time = "2026-02-14T09:34:54.179Z" }, + { url = "https://files.pythonhosted.org/packages/3f/34/3420d97065aab1b35f3e93ce9c96c8ebd423ce86fe84dee3126790421a2a/nh3-0.3.3-cp38-abi3-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:52e973cb742e95b9ae1b35822ce23992428750f4b46b619fe86eba4205255b30", size = 1029316, upload-time = "2026-02-14T09:34:56.186Z" }, + { url = "https://files.pythonhosted.org/packages/f1/9a/99eda757b14e596fdb2ca5f599a849d9554181aa899274d0d183faef4493/nh3-0.3.3-cp38-abi3-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:4c730617bdc15d7092dcc0469dc2826b914c8f874996d105b4bc3842a41c1cd9", size = 919944, upload-time = "2026-02-14T09:34:57.886Z" }, + { url = "https://files.pythonhosted.org/packages/6f/84/c0dc75c7fb596135f999e59a410d9f45bdabb989f1cb911f0016d22b747b/nh3-0.3.3-cp38-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e98fa3dbfd54e25487e36ba500bc29bca3a4cab4ffba18cfb1a35a2d02624297", size = 811461, upload-time = "2026-02-14T09:34:59.65Z" }, + { url = "https://files.pythonhosted.org/packages/7e/ec/b1bf57cab6230eec910e4863528dc51dcf21b57aaf7c88ee9190d62c9185/nh3-0.3.3-cp38-abi3-manylinux_2_31_riscv64.whl", hash = "sha256:3a62b8ae7c235481715055222e54c682422d0495a5c73326807d4e44c5d14691", size = 840360, upload-time = "2026-02-14T09:35:01.444Z" }, + { url = "https://files.pythonhosted.org/packages/37/5e/326ae34e904dde09af1de51219a611ae914111f0970f2f111f4f0188f57e/nh3-0.3.3-cp38-abi3-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:fc305a2264868ec8fa16548296f803d8fd9c1fa66cd28b88b605b1bd06667c0b", size = 859872, upload-time = "2026-02-14T09:35:03.348Z" }, + { url = "https://files.pythonhosted.org/packages/09/38/7eba529ce17ab4d3790205da37deabb4cb6edcba15f27b8562e467f2fc97/nh3-0.3.3-cp38-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:90126a834c18af03bfd6ff9a027bfa6bbf0e238527bc780a24de6bd7cc1041e2", size = 1023550, upload-time = "2026-02-14T09:35:04.829Z" }, + { url = "https://files.pythonhosted.org/packages/05/a2/556fdecd37c3681b1edee2cf795a6799c6ed0a5551b2822636960d7e7651/nh3-0.3.3-cp38-abi3-musllinux_1_2_armv7l.whl", hash = "sha256:24769a428e9e971e4ccfb24628f83aaa7dc3c8b41b130c8ddc1835fa1c924489", size = 1105212, upload-time = "2026-02-14T09:35:06.821Z" }, + { url = "https://files.pythonhosted.org/packages/dd/e3/5db0b0ad663234967d83702277094687baf7c498831a2d3ad3451c11770f/nh3-0.3.3-cp38-abi3-musllinux_1_2_i686.whl", hash = "sha256:b7a18ee057761e455d58b9d31445c3e4b2594cff4ddb84d2e331c011ef46f462", size = 1069970, upload-time = "2026-02-14T09:35:08.504Z" }, + { url = "https://files.pythonhosted.org/packages/79/b2/2ea21b79c6e869581ce5f51549b6e185c4762233591455bf2a326fb07f3b/nh3-0.3.3-cp38-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:5a4b2c1f3e6f3cbe7048e17f4fefad3f8d3e14cc0fd08fb8599e0d5653f6b181", size = 1047588, upload-time = "2026-02-14T09:35:09.911Z" }, + { url = "https://files.pythonhosted.org/packages/e2/92/2e434619e658c806d9c096eed2cdff9a883084299b7b19a3f0824eb8e63d/nh3-0.3.3-cp38-abi3-win32.whl", hash = "sha256:e974850b131fdffa75e7ad8e0d9c7a855b96227b093417fdf1bd61656e530f37", size = 616179, upload-time = "2026-02-14T09:35:11.366Z" }, + { url = "https://files.pythonhosted.org/packages/73/88/1ce287ef8649dc51365b5094bd3713b76454838140a32ab4f8349973883c/nh3-0.3.3-cp38-abi3-win_amd64.whl", hash = "sha256:2efd17c0355d04d39e6d79122b42662277ac10a17ea48831d90b46e5ef7e4fc0", size = 631159, upload-time = "2026-02-14T09:35:12.77Z" }, + { url = "https://files.pythonhosted.org/packages/31/f1/b4835dbde4fb06f29db89db027576d6014081cd278d9b6751facc3e69e43/nh3-0.3.3-cp38-abi3-win_arm64.whl", hash = "sha256:b838e619f483531483d26d889438e53a880510e832d2aafe73f93b7b1ac2bce2", size = 616645, upload-time = "2026-02-14T09:35:14.062Z" }, ] [[package]] @@ -5175,7 +5179,7 @@ wheels = [ [[package]] name = "requests" -version = "2.33.0" +version = "2.32.5" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "certifi" }, @@ -5183,9 +5187,9 @@ dependencies = [ { name = "idna" }, { name = "urllib3" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/34/64/8860370b167a9721e8956ae116825caff829224fbca0ca6e7bf8ddef8430/requests-2.33.0.tar.gz", hash = "sha256:c7ebc5e8b0f21837386ad0e1c8fe8b829fa5f544d8df3b2253bff14ef29d7652", size = 134232, upload-time = "2026-03-25T15:10:41.586Z" } +sdist = { url = "https://files.pythonhosted.org/packages/c9/74/b3ff8e6c8446842c3f5c837e9c3dfcfe2018ea6ecef224c710c85ef728f4/requests-2.32.5.tar.gz", hash = "sha256:dbba0bac56e100853db0ea71b82b4dfd5fe2bf6d3754a8893c3af500cec7d7cf", size = 134517, upload-time = "2025-08-18T20:46:02.573Z" } wheels = [ - { url = "https://files.pythonhosted.org/packages/56/5d/c814546c2333ceea4ba42262d8c4d55763003e767fa169adc693bd524478/requests-2.33.0-py3-none-any.whl", hash = "sha256:3324635456fa185245e24865e810cecec7b4caf933d7eb133dcde67d48cee69b", size = 65017, upload-time = "2026-03-25T15:10:40.382Z" }, + { url = "https://files.pythonhosted.org/packages/1e/db/4254e3eabe8020b458f1a747140d32277ec7a271daf1d235b70dc0b4e6e3/requests-2.32.5-py3-none-any.whl", hash = "sha256:2462f94637a34fd532264295e186976db0f5d453d1cdd31473c85a6a161affb6", size = 64738, upload-time = "2025-08-18T20:46:00.542Z" }, ] [[package]] @@ -6120,16 +6124,16 @@ wheels = [ [[package]] name = "types-boto3" -version = "1.42.75" +version = "1.42.74" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "botocore-stubs" }, { name = "types-s3transfer" }, { name = "typing-extensions", marker = "python_full_version < '3.12'" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/ed/e0/f4cb0e91d088bb96f5e4ed18b0201d802047ed69d8d62561eb7b81bd3881/types_boto3-1.42.75.tar.gz", hash = "sha256:2a41afd0cbc31171fdf93215d46fa9cfa450dc52b6775bccfbac633ce30d8617", size = 101716, upload-time = "2026-03-24T21:54:22.815Z" } +sdist = { url = "https://files.pythonhosted.org/packages/66/d2/5f1c3dfd75993084c8cf5d31e169c342fb394d7905d6438428245661c39b/types_boto3-1.42.74.tar.gz", hash = "sha256:8013a2dfc1ba398217d2d2dc6b54b37494df65cbae363f9430af4824697cb655", size = 101691, upload-time = "2026-03-23T19:56:24.402Z" } wheels = [ - { url = "https://files.pythonhosted.org/packages/ec/d5/4477cfd9e3f41b4dd7492b998c80985e279926ad06686fd169b77b78a6e1/types_boto3-1.42.75-py3-none-any.whl", hash = "sha256:c285c4d8e9358ed95feff5fe108847e52223ae344cc24699db29a4ec4fa37b43", size = 69908, upload-time = "2026-03-24T21:54:09.684Z" }, + { url = "https://files.pythonhosted.org/packages/6c/b1/8836349f6c9a3bf2066fb8eaec74180406f0a11e8ce00fc33232f97a8c38/types_boto3-1.42.74-py3-none-any.whl", hash = "sha256:3791c49c694b5c6d980e38994032eff4ed90ab4f7b1ad4fb34f3501b7fc60d02", size = 69907, upload-time = "2026-03-23T19:56:19.283Z" }, ] [package.optional-dependencies] @@ -6139,14 +6143,14 @@ full = [ [[package]] name = "types-boto3-full" -version = "1.42.75" +version = "1.42.74" source = { registry = "https://pypi.org/simple" } dependencies = [ { name = "typing-extensions", marker = "python_full_version < '3.12'" }, ] -sdist = { url = "https://files.pythonhosted.org/packages/d0/96/e69885de1e3018b3319fd65e537cb9adf30fc56ad2a72ecff4d86293f737/types_boto3_full-1.42.75.tar.gz", hash = "sha256:3c1ca5f6c153aef801449717f21f37c79a7e7a08c7fc744a606600c4ab8b0dbe", size = 8500715, upload-time = "2026-03-25T01:45:21.413Z" } +sdist = { url = "https://files.pythonhosted.org/packages/c9/4d/5fa22d7c99b57e9f92c522b10d96caab0dc638880863cc86901e7f9316da/types_boto3_full-1.42.74.tar.gz", hash = "sha256:827c31e18c1cc8720b826b3c3b8f7be3612c5fc3800d61d0591e9d65257ecb0a", size = 8499614, upload-time = "2026-03-24T01:27:55.394Z" } wheels = [ - { url = "https://files.pythonhosted.org/packages/d4/9a/30c19c379728d835674b47842a0620e566eb661fc345e45d9b8f87656c68/types_boto3_full-1.42.75-py3-none-any.whl", hash = "sha256:95d8e378d67f032fea1440fbb90361c68a47fdf41e8c1ffd86539375d6d845ea", size = 12908515, upload-time = "2026-03-25T01:45:17.202Z" }, + { url = "https://files.pythonhosted.org/packages/34/c0/9b55d89c3ec01aa91518cae52b73a03af6ad8c07e46551d865f40039fc23/types_boto3_full-1.42.74-py3-none-any.whl", hash = "sha256:654f814175af53caa1c34498883c6edcdb8555066e19bbe800f98d1d8b862d29", size = 12908188, upload-time = "2026-03-24T01:27:51.64Z" }, ] [[package]]