orchestrator-core 4.4.0rc1__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 (65) 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/app.py +4 -0
  7. orchestrator/cli/index_llm.py +73 -0
  8. orchestrator/cli/main.py +8 -1
  9. orchestrator/cli/resize_embedding.py +136 -0
  10. orchestrator/cli/scheduler.py +29 -39
  11. orchestrator/cli/search_explore.py +203 -0
  12. orchestrator/db/models.py +37 -1
  13. orchestrator/graphql/schema.py +0 -5
  14. orchestrator/graphql/schemas/process.py +2 -2
  15. orchestrator/graphql/utils/create_resolver_error_handler.py +1 -1
  16. orchestrator/migrations/versions/schema/2025-08-12_52b37b5b2714_search_index_model_for_llm_integration.py +95 -0
  17. orchestrator/schedules/__init__.py +2 -1
  18. orchestrator/schedules/resume_workflows.py +2 -2
  19. orchestrator/schedules/scheduling.py +24 -64
  20. orchestrator/schedules/task_vacuum.py +2 -2
  21. orchestrator/schedules/validate_products.py +2 -8
  22. orchestrator/schedules/validate_subscriptions.py +2 -2
  23. orchestrator/schemas/search.py +101 -0
  24. orchestrator/search/__init__.py +0 -0
  25. orchestrator/search/agent/__init__.py +1 -0
  26. orchestrator/search/agent/prompts.py +62 -0
  27. orchestrator/search/agent/state.py +8 -0
  28. orchestrator/search/agent/tools.py +122 -0
  29. orchestrator/search/core/__init__.py +0 -0
  30. orchestrator/search/core/embedding.py +64 -0
  31. orchestrator/search/core/exceptions.py +16 -0
  32. orchestrator/search/core/types.py +162 -0
  33. orchestrator/search/core/validators.py +27 -0
  34. orchestrator/search/docs/index.md +37 -0
  35. orchestrator/search/docs/running_local_text_embedding_inference.md +45 -0
  36. orchestrator/search/filters/__init__.py +27 -0
  37. orchestrator/search/filters/base.py +236 -0
  38. orchestrator/search/filters/date_filters.py +75 -0
  39. orchestrator/search/filters/definitions.py +76 -0
  40. orchestrator/search/filters/ltree_filters.py +31 -0
  41. orchestrator/search/filters/numeric_filter.py +60 -0
  42. orchestrator/search/indexing/__init__.py +3 -0
  43. orchestrator/search/indexing/indexer.py +316 -0
  44. orchestrator/search/indexing/registry.py +88 -0
  45. orchestrator/search/indexing/tasks.py +53 -0
  46. orchestrator/search/indexing/traverse.py +209 -0
  47. orchestrator/search/retrieval/__init__.py +3 -0
  48. orchestrator/search/retrieval/builder.py +64 -0
  49. orchestrator/search/retrieval/engine.py +96 -0
  50. orchestrator/search/retrieval/ranker.py +202 -0
  51. orchestrator/search/retrieval/utils.py +88 -0
  52. orchestrator/search/retrieval/validation.py +174 -0
  53. orchestrator/search/schemas/__init__.py +0 -0
  54. orchestrator/search/schemas/parameters.py +114 -0
  55. orchestrator/search/schemas/results.py +47 -0
  56. orchestrator/services/processes.py +11 -16
  57. orchestrator/settings.py +29 -1
  58. orchestrator/workflow.py +1 -8
  59. {orchestrator_core-4.4.0rc1.dist-info → orchestrator_core-5.0.0a1.dist-info}/METADATA +6 -3
  60. {orchestrator_core-4.4.0rc1.dist-info → orchestrator_core-5.0.0a1.dist-info}/RECORD +62 -26
  61. orchestrator/graphql/resolvers/scheduled_tasks.py +0 -36
  62. orchestrator/graphql/schemas/scheduled_task.py +0 -8
  63. orchestrator/schedules/scheduler.py +0 -153
  64. {orchestrator_core-4.4.0rc1.dist-info → orchestrator_core-5.0.0a1.dist-info}/WHEEL +0 -0
  65. {orchestrator_core-4.4.0rc1.dist-info → orchestrator_core-5.0.0a1.dist-info}/licenses/LICENSE +0 -0
@@ -1,153 +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 datetime import datetime
16
- from typing import Any
17
-
18
- from apscheduler.jobstores.sqlalchemy import SQLAlchemyJobStore
19
- from apscheduler.schedulers.background import BackgroundScheduler
20
- from more_itertools import partition
21
- from pydantic import BaseModel
22
-
23
- from orchestrator.db.filters import Filter
24
- from orchestrator.db.filters.filters import CallableErrorHandler
25
- from orchestrator.db.sorting import Sort
26
- from orchestrator.db.sorting.sorting import SortOrder
27
- from orchestrator.settings import app_settings
28
- from orchestrator.utils.helpers import camel_to_snake, to_camel
29
-
30
- jobstores = {"default": SQLAlchemyJobStore(url=str(app_settings.DATABASE_URI))}
31
-
32
- scheduler = BackgroundScheduler(jobstores=jobstores)
33
-
34
-
35
- def scheduler_dispose_db_connections() -> None:
36
- jobstores["default"].engine.dispose()
37
-
38
-
39
- class ScheduledTask(BaseModel):
40
- id: str
41
- name: str | None = None
42
- next_run_time: datetime | None = None
43
- trigger: str
44
-
45
-
46
- scheduled_task_keys = set(ScheduledTask.model_fields.keys())
47
- scheduled_task_filter_keys = sorted(scheduled_task_keys | {to_camel(key) for key in scheduled_task_keys})
48
- scheduled_task_sort_keys = scheduled_task_filter_keys
49
-
50
-
51
- def scheduled_task_in_filter(job: ScheduledTask, filter_by: list[Filter]) -> bool:
52
- return any(f.value.lower() in getattr(job, camel_to_snake(f.field), "").lower() for f in filter_by)
53
-
54
-
55
- def filter_scheduled_tasks(
56
- scheduled_tasks: list[ScheduledTask],
57
- handle_filter_error: CallableErrorHandler,
58
- filter_by: list[Filter] | None = None,
59
- ) -> list[ScheduledTask]:
60
- if not filter_by:
61
- return scheduled_tasks
62
-
63
- try:
64
- invalid_filters, valid_filters = partition(lambda x: x.field in scheduled_task_filter_keys, filter_by)
65
-
66
- if invalid_list := [item.field for item in invalid_filters]:
67
- handle_filter_error(
68
- "Invalid filter arguments", invalid_filters=invalid_list, valid_filter_keys=scheduled_task_filter_keys
69
- )
70
-
71
- valid_filter_list = list(valid_filters)
72
- return [task for task in scheduled_tasks if scheduled_task_in_filter(task, valid_filter_list)]
73
- except Exception as e:
74
- handle_filter_error(str(e))
75
- return []
76
-
77
-
78
- def _invert(value: Any) -> Any:
79
- """Invert value for descending order."""
80
- if isinstance(value, (int, float)):
81
- return -value
82
- if isinstance(value, str):
83
- return tuple(-ord(c) for c in value)
84
- if isinstance(value, datetime):
85
- return -value.timestamp()
86
- return value
87
-
88
-
89
- def sort_key(sort_field: str, sort_order: SortOrder) -> Any:
90
- def _sort_key(task: Any) -> Any:
91
- value = getattr(task, camel_to_snake(sort_field), None)
92
- if sort_field == "next_run_time" and value is None:
93
- return float("inf") if sort_order == SortOrder.ASC else float("-inf")
94
- return value if sort_order == SortOrder.ASC else _invert(value)
95
-
96
- return _sort_key
97
-
98
-
99
- def sort_scheduled_tasks(
100
- scheduled_tasks: list[ScheduledTask], handle_sort_error: CallableErrorHandler, sort_by: list[Sort] | None = None
101
- ) -> list[ScheduledTask]:
102
- if not sort_by:
103
- return scheduled_tasks
104
-
105
- try:
106
- invalid_sorting, valid_sorting = partition(lambda x: x.field in scheduled_task_sort_keys, sort_by)
107
- if invalid_list := [item.field for item in invalid_sorting]:
108
- handle_sort_error(
109
- "Invalid sort arguments", invalid_sorting=invalid_list, valid_sort_keys=scheduled_task_sort_keys
110
- )
111
-
112
- valid_sort_list = list(valid_sorting)
113
- return sorted(
114
- scheduled_tasks, key=lambda task: tuple(sort_key(sort.field, sort.order)(task) for sort in valid_sort_list)
115
- )
116
- except Exception as e:
117
- handle_sort_error(str(e))
118
- return []
119
-
120
-
121
- def default_error_handler(message: str, **context) -> None: # type: ignore
122
- from orchestrator.graphql.utils.create_resolver_error_handler import _format_context
123
-
124
- raise ValueError(f"{message} {_format_context(context)}")
125
-
126
-
127
- def get_scheduler_tasks(
128
- first: int = 10,
129
- after: int = 0,
130
- filter_by: list[Filter] | None = None,
131
- sort_by: list[Sort] | None = None,
132
- error_handler: CallableErrorHandler = default_error_handler,
133
- ) -> tuple[list[ScheduledTask], int]:
134
- scheduler.start(paused=True)
135
- scheduled_tasks = scheduler.get_jobs()
136
- scheduler.shutdown(wait=False)
137
- scheduler_dispose_db_connections()
138
-
139
- scheduled_tasks = filter_scheduled_tasks(scheduled_tasks, error_handler, filter_by)
140
- scheduled_tasks = sort_scheduled_tasks(scheduled_tasks, error_handler, sort_by)
141
-
142
- total = len(scheduled_tasks)
143
- paginated_tasks = scheduled_tasks[after : after + first + 1]
144
-
145
- return [
146
- ScheduledTask(
147
- id=task.id,
148
- name=task.name,
149
- next_run_time=task.next_run_time,
150
- trigger=str(task.trigger),
151
- )
152
- for task in paginated_tasks
153
- ], total