orchestrator-core 4.4.0rc2__py3-none-any.whl → 5.0.0a1__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (69) hide show
  1. orchestrator/__init__.py +1 -1
  2. orchestrator/api/api_v1/api.py +7 -0
  3. orchestrator/api/api_v1/endpoints/agent.py +62 -0
  4. orchestrator/api/api_v1/endpoints/processes.py +6 -12
  5. orchestrator/api/api_v1/endpoints/search.py +197 -0
  6. orchestrator/api/api_v1/endpoints/subscriptions.py +0 -1
  7. orchestrator/app.py +4 -0
  8. orchestrator/cli/index_llm.py +73 -0
  9. orchestrator/cli/main.py +8 -1
  10. orchestrator/cli/resize_embedding.py +136 -0
  11. orchestrator/cli/scheduler.py +29 -40
  12. orchestrator/cli/search_explore.py +203 -0
  13. orchestrator/db/models.py +37 -1
  14. orchestrator/graphql/schema.py +0 -5
  15. orchestrator/graphql/schemas/process.py +2 -2
  16. orchestrator/graphql/utils/create_resolver_error_handler.py +1 -1
  17. orchestrator/migrations/versions/schema/2025-08-12_52b37b5b2714_search_index_model_for_llm_integration.py +95 -0
  18. orchestrator/schedules/__init__.py +2 -1
  19. orchestrator/schedules/resume_workflows.py +2 -2
  20. orchestrator/schedules/scheduling.py +24 -64
  21. orchestrator/schedules/task_vacuum.py +2 -2
  22. orchestrator/schedules/validate_products.py +2 -8
  23. orchestrator/schedules/validate_subscriptions.py +2 -2
  24. orchestrator/schemas/search.py +101 -0
  25. orchestrator/search/__init__.py +0 -0
  26. orchestrator/search/agent/__init__.py +1 -0
  27. orchestrator/search/agent/prompts.py +62 -0
  28. orchestrator/search/agent/state.py +8 -0
  29. orchestrator/search/agent/tools.py +122 -0
  30. orchestrator/search/core/__init__.py +0 -0
  31. orchestrator/search/core/embedding.py +64 -0
  32. orchestrator/search/core/exceptions.py +16 -0
  33. orchestrator/search/core/types.py +162 -0
  34. orchestrator/search/core/validators.py +27 -0
  35. orchestrator/search/docs/index.md +37 -0
  36. orchestrator/search/docs/running_local_text_embedding_inference.md +45 -0
  37. orchestrator/search/filters/__init__.py +27 -0
  38. orchestrator/search/filters/base.py +236 -0
  39. orchestrator/search/filters/date_filters.py +75 -0
  40. orchestrator/search/filters/definitions.py +76 -0
  41. orchestrator/search/filters/ltree_filters.py +31 -0
  42. orchestrator/search/filters/numeric_filter.py +60 -0
  43. orchestrator/search/indexing/__init__.py +3 -0
  44. orchestrator/search/indexing/indexer.py +316 -0
  45. orchestrator/search/indexing/registry.py +88 -0
  46. orchestrator/search/indexing/tasks.py +53 -0
  47. orchestrator/search/indexing/traverse.py +209 -0
  48. orchestrator/search/retrieval/__init__.py +3 -0
  49. orchestrator/search/retrieval/builder.py +64 -0
  50. orchestrator/search/retrieval/engine.py +96 -0
  51. orchestrator/search/retrieval/ranker.py +202 -0
  52. orchestrator/search/retrieval/utils.py +88 -0
  53. orchestrator/search/retrieval/validation.py +174 -0
  54. orchestrator/search/schemas/__init__.py +0 -0
  55. orchestrator/search/schemas/parameters.py +114 -0
  56. orchestrator/search/schemas/results.py +47 -0
  57. orchestrator/services/processes.py +11 -16
  58. orchestrator/services/subscriptions.py +0 -4
  59. orchestrator/settings.py +29 -1
  60. orchestrator/targets.py +0 -1
  61. orchestrator/workflow.py +1 -8
  62. orchestrator/workflows/utils.py +1 -48
  63. {orchestrator_core-4.4.0rc2.dist-info → orchestrator_core-5.0.0a1.dist-info}/METADATA +6 -3
  64. {orchestrator_core-4.4.0rc2.dist-info → orchestrator_core-5.0.0a1.dist-info}/RECORD +66 -30
  65. orchestrator/graphql/resolvers/scheduled_tasks.py +0 -36
  66. orchestrator/graphql/schemas/scheduled_task.py +0 -8
  67. orchestrator/schedules/scheduler.py +0 -163
  68. {orchestrator_core-4.4.0rc2.dist-info → orchestrator_core-5.0.0a1.dist-info}/WHEEL +0 -0
  69. {orchestrator_core-4.4.0rc2.dist-info → orchestrator_core-5.0.0a1.dist-info}/licenses/LICENSE +0 -0
@@ -1,36 +0,0 @@
1
- import structlog
2
-
3
- from orchestrator.db.filters import Filter
4
- from orchestrator.db.sorting import Sort
5
- from orchestrator.graphql.pagination import Connection
6
- from orchestrator.graphql.schemas.scheduled_task import ScheduledTaskGraphql
7
- from orchestrator.graphql.types import GraphqlFilter, GraphqlSort, OrchestratorInfo
8
- from orchestrator.graphql.utils import create_resolver_error_handler, to_graphql_result_page
9
- from orchestrator.graphql.utils.is_query_detailed import is_querying_page_data
10
- from orchestrator.schedules.scheduler import get_scheduler_tasks, scheduled_task_filter_keys, scheduled_task_sort_keys
11
-
12
- logger = structlog.get_logger(__name__)
13
-
14
-
15
- async def resolve_scheduled_tasks(
16
- info: OrchestratorInfo,
17
- filter_by: list[GraphqlFilter] | None = None,
18
- sort_by: list[GraphqlSort] | None = None,
19
- first: int = 10,
20
- after: int = 0,
21
- ) -> Connection[ScheduledTaskGraphql]:
22
- _error_handler = create_resolver_error_handler(info)
23
-
24
- pydantic_filter_by: list[Filter] = [item.to_pydantic() for item in filter_by] if filter_by else []
25
- pydantic_sort_by: list[Sort] = [item.to_pydantic() for item in sort_by] if sort_by else []
26
- scheduled_tasks, total = get_scheduler_tasks(
27
- first=first, after=after, filter_by=pydantic_filter_by, sort_by=pydantic_sort_by, error_handler=_error_handler
28
- )
29
-
30
- graphql_scheduled_tasks = []
31
- if is_querying_page_data(info):
32
- graphql_scheduled_tasks = [ScheduledTaskGraphql.from_pydantic(p) for p in scheduled_tasks]
33
-
34
- return to_graphql_result_page(
35
- graphql_scheduled_tasks, first, after, total, scheduled_task_filter_keys, scheduled_task_sort_keys
36
- )
@@ -1,8 +0,0 @@
1
- import strawberry
2
-
3
- from orchestrator.schedules.scheduler import ScheduledTask
4
-
5
-
6
- @strawberry.experimental.pydantic.type(model=ScheduledTask, all_fields=True)
7
- class ScheduledTaskGraphql:
8
- pass
@@ -1,163 +0,0 @@
1
- # Copyright 2019-2020 SURF.
2
- # Licensed under the Apache License, Version 2.0 (the "License");
3
- # you may not use this file except in compliance with the License.
4
- # You may obtain a copy of the License at
5
- #
6
- # http://www.apache.org/licenses/LICENSE-2.0
7
- #
8
- # Unless required by applicable law or agreed to in writing, software
9
- # distributed under the License is distributed on an "AS IS" BASIS,
10
- # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
11
- # See the License for the specific language governing permissions and
12
- # limitations under the License.
13
-
14
-
15
- from contextlib import contextmanager
16
- from datetime import datetime
17
- from typing import Any, Generator
18
-
19
- from apscheduler.jobstores.sqlalchemy import SQLAlchemyJobStore
20
- from apscheduler.schedulers.background import BackgroundScheduler
21
- from more_itertools import partition
22
- from pydantic import BaseModel
23
-
24
- from orchestrator.db.filters import Filter
25
- from orchestrator.db.filters.filters import CallableErrorHandler
26
- from orchestrator.db.sorting import Sort
27
- from orchestrator.db.sorting.sorting import SortOrder
28
- from orchestrator.settings import app_settings
29
- from orchestrator.utils.helpers import camel_to_snake, to_camel
30
-
31
- jobstores = {"default": SQLAlchemyJobStore(url=str(app_settings.DATABASE_URI))}
32
-
33
- scheduler = BackgroundScheduler(jobstores=jobstores)
34
-
35
-
36
- def scheduler_dispose_db_connections() -> None:
37
- jobstores["default"].engine.dispose()
38
-
39
-
40
- @contextmanager
41
- def get_paused_scheduler() -> Generator[BackgroundScheduler, Any, None]:
42
- scheduler.start(paused=True)
43
-
44
- try:
45
- yield scheduler
46
- finally:
47
- scheduler.shutdown(wait=False)
48
- scheduler_dispose_db_connections()
49
-
50
-
51
- class ScheduledTask(BaseModel):
52
- id: str
53
- name: str | None = None
54
- next_run_time: datetime | None = None
55
- trigger: str
56
-
57
-
58
- scheduled_task_keys = set(ScheduledTask.model_fields.keys())
59
- scheduled_task_filter_keys = sorted(scheduled_task_keys | {to_camel(key) for key in scheduled_task_keys})
60
- scheduled_task_sort_keys = scheduled_task_filter_keys
61
-
62
-
63
- def scheduled_task_in_filter(job: ScheduledTask, filter_by: list[Filter]) -> bool:
64
- return any(f.value.lower() in getattr(job, camel_to_snake(f.field), "").lower() for f in filter_by)
65
-
66
-
67
- def filter_scheduled_tasks(
68
- scheduled_tasks: list[ScheduledTask],
69
- handle_filter_error: CallableErrorHandler,
70
- filter_by: list[Filter] | None = None,
71
- ) -> list[ScheduledTask]:
72
- if not filter_by:
73
- return scheduled_tasks
74
-
75
- try:
76
- invalid_filters, valid_filters = partition(lambda x: x.field in scheduled_task_filter_keys, filter_by)
77
-
78
- if invalid_list := [item.field for item in invalid_filters]:
79
- handle_filter_error(
80
- "Invalid filter arguments", invalid_filters=invalid_list, valid_filter_keys=scheduled_task_filter_keys
81
- )
82
-
83
- valid_filter_list = list(valid_filters)
84
- return [task for task in scheduled_tasks if scheduled_task_in_filter(task, valid_filter_list)]
85
- except Exception as e:
86
- handle_filter_error(str(e))
87
- return []
88
-
89
-
90
- def _invert(value: Any) -> Any:
91
- """Invert value for descending order."""
92
- if isinstance(value, (int, float)):
93
- return -value
94
- if isinstance(value, str):
95
- return tuple(-ord(c) for c in value)
96
- if isinstance(value, datetime):
97
- return -value.timestamp()
98
- return value
99
-
100
-
101
- def sort_key(sort_field: str, sort_order: SortOrder) -> Any:
102
- def _sort_key(task: Any) -> Any:
103
- value = getattr(task, camel_to_snake(sort_field), None)
104
- if sort_field == "next_run_time" and value is None:
105
- return float("inf") if sort_order == SortOrder.ASC else float("-inf")
106
- return value if sort_order == SortOrder.ASC else _invert(value)
107
-
108
- return _sort_key
109
-
110
-
111
- def sort_scheduled_tasks(
112
- scheduled_tasks: list[ScheduledTask], handle_sort_error: CallableErrorHandler, sort_by: list[Sort] | None = None
113
- ) -> list[ScheduledTask]:
114
- if not sort_by:
115
- return scheduled_tasks
116
-
117
- try:
118
- invalid_sorting, valid_sorting = partition(lambda x: x.field in scheduled_task_sort_keys, sort_by)
119
- if invalid_list := [item.field for item in invalid_sorting]:
120
- handle_sort_error(
121
- "Invalid sort arguments", invalid_sorting=invalid_list, valid_sort_keys=scheduled_task_sort_keys
122
- )
123
-
124
- valid_sort_list = list(valid_sorting)
125
- return sorted(
126
- scheduled_tasks, key=lambda task: tuple(sort_key(sort.field, sort.order)(task) for sort in valid_sort_list)
127
- )
128
- except Exception as e:
129
- handle_sort_error(str(e))
130
- return []
131
-
132
-
133
- def default_error_handler(message: str, **context) -> None: # type: ignore
134
- from orchestrator.graphql.utils.create_resolver_error_handler import _format_context
135
-
136
- raise ValueError(f"{message} {_format_context(context)}")
137
-
138
-
139
- def get_scheduler_tasks(
140
- first: int = 10,
141
- after: int = 0,
142
- filter_by: list[Filter] | None = None,
143
- sort_by: list[Sort] | None = None,
144
- error_handler: CallableErrorHandler = default_error_handler,
145
- ) -> tuple[list[ScheduledTask], int]:
146
- with get_paused_scheduler() as pauzed_scheduler:
147
- scheduled_tasks = pauzed_scheduler.get_jobs()
148
-
149
- scheduled_tasks = filter_scheduled_tasks(scheduled_tasks, error_handler, filter_by)
150
- scheduled_tasks = sort_scheduled_tasks(scheduled_tasks, error_handler, sort_by)
151
-
152
- total = len(scheduled_tasks)
153
- paginated_tasks = scheduled_tasks[after : after + first + 1]
154
-
155
- return [
156
- ScheduledTask(
157
- id=task.id,
158
- name=task.name,
159
- next_run_time=task.next_run_time,
160
- trigger=str(task.trigger),
161
- )
162
- for task in paginated_tasks
163
- ], total