by-qa 0.4.4.dev4__tar.gz → 0.4.4.dev5__tar.gz

{by_qa-0.4.4.dev4 → by_qa-0.4.4.dev5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: by-qa
-Version: 0.4.4.dev4
+Version: 0.4.4.dev5
 Summary: Open-source knowledge-base ingestion and retrieval service.
 Author: AI Assistant
 License: MIT

{by_qa-0.4.4.dev4 → by_qa-0.4.4.dev5}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "by-qa"
-version = "0.4.4.dev4"
+version = "0.4.4.dev5"
 description = "Open-source knowledge-base ingestion and retrieval service."
 readme = "README.md"
 requires-python = ">=3.12"

{by_qa-0.4.4.dev4 → by_qa-0.4.4.dev5}/src/by_qa/knowledge_base/api/routes.py

@@ -44,6 +44,7 @@ from by_qa.knowledge_base.api.schemas import (
     FileBuildStatusRequest,
     FileToMarkdownIndexRequest,
     KnowledgeBaseFileCountRequest,
+    KnowledgeBaseStatusRequest,
     KnowledgeItemDownloadRequest,
     KnowledgeItemGlobRequest,
     KnowledgeItemListDirRequest,
@@ -512,6 +513,11 @@ def register_routes(
 
     @app.post(
         "/api/v1/knowledgeBases/fileCount",
+        deprecated=True,
+        description=(
+            "Deprecated. Use POST /api/v1/knowledgeBases/status with knCodeList "
+            "to retrieve fileCount and spaceSize for each knowledge base."
+        ),
         openapi_extra={
             "requestBody": {
                 "required": True,
@@ -565,6 +571,62 @@ def register_routes(
             )
         return _documented_success_response(result_object=result)
 
+    @app.post(
+        "/api/v1/knowledgeBases/status",
+        openapi_extra={
+            "requestBody": {
+                "required": True,
+                "content": {
+                    "application/json": {
+                        "schema": KnowledgeBaseStatusRequest.model_json_schema(
+                            by_alias=True
+                        )
+                    }
+                },
+            }
+        },
+    )
+    async def knowledge_base_status(http_request: Request):
+        # Parse locally so malformed bodies also honor this endpoint's HTTP 422
+        # contract rather than the application's legacy HTTP 200 error handler.
+        try:
+            request = KnowledgeBaseStatusRequest.model_validate(
+                await http_request.json()
+            )
+        except ValueError:
+            return _documented_error_response(
+                result_msg="request validation failed",
+                status_code=422,
+                preserve_http_status=True,
+            )
+
+        logger.info(
+            "knowledge_base_status request received: kb_code_list=%s",
+            request.kb_code_list,
+        )
+        try:
+            service = await get_knowledge_base_service()
+            result = await service.status(request)
+        except KnowledgeBaseConfigurationError as exc:
+            return _documented_error_response(
+                result_msg=str(exc), status_code=503, preserve_http_status=True
+            )
+        except KnowledgeBaseValidationError as exc:
+            return _documented_error_response(
+                result_msg=str(exc), status_code=422, preserve_http_status=True
+            )
+        except Exception as exc:
+            logger.exception(
+                "knowledge_base_status unexpected error: kb_code_list=%s",
+                request.kb_code_list,
+            )
+            return _documented_error_response(
+                result_msg=str(exc) or "internal error",
+                status_code=500,
+                preserve_http_status=True,
+            )
+        return _documented_success_response(result_object=result)
+
     @app.post("/api/v1/directories/create")
     async def create_directory(
         background_tasks: BackgroundTasks,

{by_qa-0.4.4.dev4 → by_qa-0.4.4.dev5}/src/by_qa/knowledge_base/api/schemas.py

@@ -1,8 +1,16 @@
 """Pydantic schemas for knowledge base APIs."""
 
-from typing import Any, Literal, Optional
+from typing import Annotated, Any, Literal, Optional
 
-from pydantic import AliasChoices, BaseModel, ConfigDict, Field, model_validator
+from pydantic import (
+    AliasChoices,
+    BaseModel,
+    ConfigDict,
+    Field,
+    StringConstraints,
+    field_validator,
+    model_validator,
+)
 
 
 def _normalize_api_path(path: str, *, allow_root: bool) -> str:
@@ -65,6 +73,27 @@ class KnowledgeBaseFileCountRequest(BaseModel):
     )
 
 
+class KnowledgeBaseStatusRequest(BaseModel):
+    """Request body for batch file counts and logical storage sizes."""
+
+    model_config = ConfigDict(populate_by_name=True)
+
+    kb_code_list: list[
+        Annotated[
+            str, StringConstraints(strict=True, strip_whitespace=True, min_length=1)
+        ]
+    ] = Field(
+        min_length=1,
+        strict=True,
+        validation_alias=AliasChoices("knCodeList", "kb_code_list"),
+    )
+
+    @field_validator("kb_code_list")
+    @classmethod
+    def deduplicate_codes(cls, codes: list[str]) -> list[str]:
+        return list(dict.fromkeys(codes))
+
+
 class UpdateKnowledgeBaseRequest(BaseModel):
     """Request body for updating knowledge base business fields."""
 

{by_qa-0.4.4.dev4 → by_qa-0.4.4.dev5}/src/by_qa/knowledge_base/repositories/knowledge_fs_entry_repository.py

@@ -1276,6 +1276,28 @@ class KnowledgeFsEntryRepository:
         row = await cursor.fetchone()
         return int(row["file_count"])
 
+    async def file_stats_by_knowledge_base_ids(
+        self, cursor: Any, *, knowledge_base_ids: list[int]
+    ) -> dict[int, dict[str, int]]:
+        """Aggregate current live file counts and byte totals in one query."""
+        await cursor.execute(
+            """SELECT knowledge_base_id, COUNT(*) AS file_count,
+                      COALESCE(SUM(COALESCE(file_size, 0)), 0) AS size_bytes
+            FROM knowledge_fs_entry
+            WHERE knowledge_base_id = ANY(%(knowledge_base_ids)s)
+              AND is_deleted = FALSE
+              AND entry_type = 'FILE'
+            GROUP BY knowledge_base_id""",
+            {"knowledge_base_ids": knowledge_base_ids},
+        )
+        return {
+            int(row["knowledge_base_id"]): {
+                "file_count": int(row["file_count"]),
+                "size_bytes": int(row["size_bytes"]),
+            }
+            for row in await self._fetchall(cursor)
+        }
+
     async def list_file_entries_by_knowledge_base_id(
         self, cursor, *, knowledge_base_id: int
     ) -> list[dict[str, Any]]:

{by_qa-0.4.4.dev4 → by_qa-0.4.4.dev5}/src/by_qa/knowledge_base/services/knowledge_base_service.py

@@ -4,7 +4,7 @@ import fnmatch
 import json
 import mimetypes
 from dataclasses import dataclass
-from decimal import Decimal
+from decimal import ROUND_HALF_UP, Decimal
 from pathlib import Path, PurePosixPath
 from typing import Any, Callable
 
@@ -20,6 +20,7 @@ from by_qa.knowledge_base.api.schemas import (
     DeleteKnowledgeBaseRequest,
     FileBuildStatusRequest,
     KnowledgeBaseFileCountRequest,
+    KnowledgeBaseStatusRequest,
     KnowledgeItemDownloadRequest,
     KnowledgeItemGlobRequest,
     KnowledgeItemListDirItem,
@@ -1012,6 +1013,39 @@ class KnowledgeBaseService:
         finally:
             await connection.close()
 
+    async def status(self, request: KnowledgeBaseStatusRequest) -> dict[str, Any]:
+        """Return ordered per-KB counts and sizes in KB, or fail the whole batch."""
+        connection = await self.connection_factory()
+        try:
+            cursor = connection.cursor()
+            kb_ids = []
+            for code in request.kb_code_list:
+                row = await self.knowledge_base_repository.get_by_code(cursor, code)
+                if not row:
+                    raise KnowledgeBaseValidationError(
+                        f"knowledge base not found: {code}"
+                    )
+                kb_ids.append(self._row_id(row))
+            stats = await self.knowledge_fs_entry_repository.file_stats_by_knowledge_base_ids(
+                cursor, knowledge_base_ids=kb_ids
+            )
+            data = []
+            for code, kb_id in zip(request.kb_code_list, kb_ids):
+                stat = stats.get(kb_id, {"file_count": 0, "size_bytes": 0})
+                size_kb = (Decimal(stat["size_bytes"]) / Decimal(1024)).quantize(
+                    Decimal("0.01"), rounding=ROUND_HALF_UP
+                )
+                data.append(
+                    {
+                        "knCode": code,
+                        "fileCount": stat["file_count"],
+                        "spaceSize": float(size_kb),
+                    }
+                )
+            return {"data": data}
+        finally:
+            await connection.close()
+
     async def list_dir(
         self, request: KnowledgeItemListDirRequest
     ) -> KnowledgeItemListDirResponse:

{by_qa-0.4.4.dev4 → by_qa-0.4.4.dev5}/uv.lock

@@ -313,7 +313,7 @@ wheels = [
 
 [[package]]
 name = "by-qa"
-version = "0.4.4.dev4"
+version = "0.4.4.dev5"
 source = { editable = "." }
 dependencies = [
     { name = "httpx" },