open-edge-platform · yangjianfeng1208 · Apr 1, 2026 · Mar 31, 2026 · Mar 31, 2026 · Mar 31, 2026
@@ -25,21 +25,18 @@ python .\start_services.py
 
 ## API Endpoints
 
-| Endpoint | Method | Pattern | Description | Status |
-| :--- | :---: | :---: | :--- | :---: |
-| `/api/v1/system/health` | **GET** | SYNC | Backend app health check | DONE |
-| `/api/v1/task/query/{task_id}` | **GET** | SYNC | Query status of a specific task | DONE |
-| `/api/v1/task/list` | **GET** | SYNC | Query tasks by conditions (e.g., `?status=PROCESSING`) | DONE |
-| `/api/v1/object/upload` | **POST** | ASYNC | Upload a file to MinIO | DONE |
-| `/api/v1/object/ingest` | **POST** | ASYNC | Ingest a specific file from MinIO | DONE |
-| `/api/v1/object/ingest-text` | **POST** | ASYNC | Emedding a raw text | DONE |
-| `/api/v1/object/upload-ingest` | **POST** | ASYNC | Upload to MinIO and trigger ingestion | DONE |
-| `/api/v1/object/search` | **POST** | ASYNC | Search for files based on description | DONE |
-| `/api/v1/object/download` | **POST** | STREAM | Download file from MinIO | DONE |
-
-## API reference
-[Content Search API reference](./docs/dev_guide/Content_search_API.md)
-
+| Endpoint | Method | Pattern | Description |
+| :--- | :---: | :---: | :--- |
+| `/api/v1/task/query/{task_id}` | **GET** | SYNC | **Task Status Inspection**: Retrieves real-time metadata for a specific job, including current lifecycle state (PENDING, PROCESSING, COMPLETED, FAILED), and error logs if applicable. |
+| `/api/v1/task/list` | **GET** | SYNC | **Batch Task Retrieval**: Queries task records. Supports filtering via query parameters (e.g., `?status=PROCESSING`) for monitoring system load and pipeline efficiency. |
+| `/api/v1/object/ingest-text` | **POST** | ASYNC | **Text-Specific Ingestion**: Primarily processes raw text strings passed in the request body for semantic indexing. It also supports fetching content from existing text-based objects in MinIO. |
+| `/api/v1/object/upload-ingest` | **POST** | ASYNC | **Atomic Upload & Ingestion**: A unified workflow that first saves the file to MinIO and then immediately initiates the ingestion pipeline. Features full content indexing and AI-driven Video Summarization for supported video formats. |
+| `/api/v1/object/search` | **POST** | SYNC | **Semantic Content Retrieval**: Executes a similarity search across vector collections using either natural language queries or base64-encoded images. Returns ranked results with associated metadata and MinIO object references. |
+| `/api/v1/object/download` | **POST** | STREAM | **Original File Download**: Securely fetches the raw source file directly from MinIO storage. Utilizes stream-bridging to pipe binary data to the client. |
+
+For detailed descriptions and examples of each endpoint, please refer to the: [Content Search API reference](./docs/dev_guide/Content_search_API.md)
+
+## Components API reference
 [Ingest and Retrieve](./docs/dev_guide/file_ingest_and_retrieve/API_GUIDE.md)
 
 [Video Preprocess](./docs/dev_guide/video_preprocess/API_GUIDE.md)

@@ -76,24 +76,21 @@ async def ingest_existing_file(
     )
 
 class IngestTextRequest(BaseModel):
-    text: str
-    bucket_name: Optional[str] = None
-    file_path: Optional[str] = None
+    text: Optional[str] = None
+    bucket_name: Optional[str] = "content-search"
+    file_key: Optional[str] = None
     meta: Dict[str, Any] = Field(default_factory=dict)
+
 @router.post("/ingest-text")
 async def ingest_raw_text(
     request: IngestTextRequest,
     background_tasks: BackgroundTasks,
     db: Session = Depends(get_db)
 ):
-    payload = request.model_dump() 
-
-    if "tags" not in payload["meta"] or payload["meta"]["tags"] is None:
-        payload["meta"]["tags"] = ["default"]
 
     result = await task_service.handle_text_ingest(
         db,
-        payload,
+        request.model_dump(), 
         background_tasks
     )
 
@@ -102,7 +99,7 @@ async def ingest_raw_text(
             "task_id": str(result["task_id"]),
             "status": result["status"]
         },
-        message="Text ingestion started"
+        message="Text ingestion task created successfully"
     )
 
 @router.post("/upload-ingest")
@@ -145,44 +142,11 @@ async def upload_file_with_ingest(
         message="Upload and Ingest started"
     )
 
-# @router.post("/search")
-# async def file_search(payload: dict):
-#     query = payload.get("query")
-#     limit = payload.get("max_num_results", 3)
-#     if not query:
-#         raise HTTPException(status_code=400, detail="Query cannot be empty")
-
-#     search_data = await search_service.semantic_search(query, limit)
-#     return resp_200(data=search_data, message="Resource found")
-
 @router.post("/search")
-async def file_search(payload: dict):
-    query = payload.get("query")
-    image_base64 = payload.get("image_base64")
-    filters = payload.get("filter")
-    limit = payload.get("max_num_results", 10)
-
-    if not query and not image_base64:
-        raise HTTPException(status_code=400, detail="Either 'query' or 'image_base64' must be provided")
-
-    if query and image_base64:
-        raise HTTPException(status_code=400, detail="Provide only one of 'query' or 'image_base64'")
-
-    search_payload = {
-        "max_num_results": limit
-    }
-
-    if query:
-        search_payload["query"] = query
-    else:
-        search_payload["image_base64"] = image_base64
-
-    if filters:
-        search_payload["filter"] = filters
-
-    search_data = await search_service.semantic_search(search_payload)
+async def file_search(payload: dict, db: Session = Depends(get_db)):
+    result = await task_service.handle_sync_search(db, payload)
 
-    return resp_200(data=search_data, message="Search completed")
+    return resp_200(data=result, message="Search completed")
 
 @router.get("/download")
 async def download_file(file_key: str):

@@ -77,8 +77,8 @@ stateDiagram-v2
 ```
 
 ## API Endpoints
-
-### Get Task List
+### Task endpoints
+#### Get Task List
 
 * URL: /api/v1/task/list
 
@@ -142,7 +142,7 @@ Response (200 OK)
     "timestamp": 1774330753
 }
 ```
-### Task Status Polling
+#### Task Status Polling
 Used to track the progress and retrieve the final result of a submitted task.
 
 * URL: /api/v1/task/query/{task_id}
@@ -153,39 +153,50 @@ Used to track the progress and retrieve the final result of a submitted task.
 
 Request:
 ```
-curl --location 'http://127.0.0.1:9011/api/v1/task/query/56cc417c-9524-41a9-a500-9f0c44a05eac'
+curl --location 'http://127.0.0.1:9011/api/v1/task/query/6b9a6a55-d327-42fe-b05e-e0f3098fe797'
 ```
 
 Response (200 OK):
 ```json
 {
     "code": 20000,
     "data": {
-        "task_id": "e557b305-e37c-4074-a04a-ebd067efbd5d",
+        "task_id": "6b9a6a55-d327-42fe-b05e-e0f3098fe797",
         "status": "COMPLETED",
         "progress": 100,
         "result": {
-            "message": "File from MinIO successfully processed. db returns {'visual': {'insert_count': 1}}",
-            "video_summary": {
-                "type": "done",
-                "job_id": "bc6513aa-e118-4945-84a8-02922595044e",
-                "run_id": "5e405f58-03cf-4e44-9e10-85741283587a",
-                "asset_id": "classroom_8.mp4",
-                "total_chunks": 1,
-                "succeeded_chunks": 1,
-                "failed_chunks": 0,
-                "ingest_ok_chunks": 1,
-                "ingest_failed_chunks": 0,
-                "elapsed_seconds": 36.89442276954651
+            "message": "Upload only, no ingest requested",
+            "file_info": {
+                "source": "minio",
+                "file_key": "runs/9e96f16a-9689-4c25-a515-04a1040b193f/raw/text/default/phy_class.txt",
+                "bucket": "content-search",
+                "filename": "phy_class.txt",
+                "run_id": "9e96f16a-9689-4c25-a515-04a1040b193f"
             }
         }
     },
     "message": "Query successful",
-    "timestamp": 1774879431
+    "timestamp": 1774931711
 }
 ```
+### File Process
+#### File Support Matrix
+
+The system supports the following file formats for all ingestion and upload-ingest operations.
 
-### File Upload
+| Category | Supported Extensions | Processing Logic |
+| :--- | :--- | :--- |
+| **Video** | `.mp4` | Frame extraction, AI-driven summarization, and semantic indexing. |
+| **Document** | `.txt`, `.pdf`, `.docx`, `.doc`, `.pptx`, `.ppt`, `.xlsx`, `.xls` | Full-text extraction, semantic chunking, and vector embedding. |
+| **Web/Markup** | `.html`, `.htm`, `.xml`, `.md`, `.rst` | Structured text parsing and content indexing. |
+| **Image** | `.jpg`, `.png`, `.jpeg` | Visual feature embedding and similarity search indexing. |
+
+> **Technical Note**: 
+> - **Video**: Default chunking is set to 30 seconds unless the `chunk_duration` parameter is provided.
+> - **Text**: Automatic semantic segmentation is applied to ensure high-quality retrieval results.
+> - **Max File Size**: Please refer to the `CS_MAX_CONTENT_LENGTH` environment variable (Default: 100MB).
+
+#### File Upload
 Used to upload a video file and initiate an asynchronous background task.
 
 * URL: /api/v1/object/upload
@@ -205,14 +216,14 @@ Response (200 OK):
     "code": 20000,
     "data": {
         "task_id": "c68211de-2187-4f52-b47d-f3a51a52b9ca",
-        "status": "QUEUED"
+        "status": "PROCESSING"
     },
     "message": "File received, processing started.",
     "timestamp": 1773909147
 }
 ```
 
-### File ingestion
+#### File ingestion
 * URL: /api/v1/object/ingest
 * Method: POST
 * Pattern: ASYNC
@@ -248,8 +259,47 @@ Response:
     "timestamp": 1774878031
 }
 ```
+#### Text file ingestion
+Used to trigger the ingestion pipeline for text-based documents (e.g., .txt, .pdf, .docx) that already exist in MinIO.
+
+* URL: /api/v1/object/ingest-text
+* Method: POST
+* Pattern: ASYNC
+* Parameters:
+
+| Field | Type | Required | Default | Description |
+| :--- | :--- | :--- | :--- | :--- |
+| `text` | `string` | **Yes** | — | **Raw text content** to be segmented, embedded, and stored in the vector database. |
+| `bucket_name` | `string` | No | — | MinIO bucket name (used to logically group the data or build the identifier). |
+| `file_path` | `string` | No | — | Logical path or filename (used as a unique identifier for the text source). |
+| `meta` | `object` | No | `{}` | Extra metadata to store alongside the text (e.g., `course`, `author`, `tags`). |
+
+Request:
+```
+<!-- example for raw text content -->
+curl --location 'http://127.0.0.1:9011/api/v1/object/ingest-text' \
+--header 'Content-Type: application/json' \
+--data '{
+    "text": "Newton'\''s Second Law of Motion states that the force acting on an object is equal to the mass of that object multiplied by its acceleration (F = ma). This relationship describes how the velocity of an object changes when it is subjected to an external force.",
+    "meta": {
+        "source": "topic-search"
+    }
+}'
+```
+Response:
+```json
+{
+    "code": 20000,
+    "data": {
+        "task_id": "f5eb96fd-9c75-4dee-a715-4d39b0762436",
+        "status": "PROCESSING"
+    },
+    "message": "Text ingestion task created successfully",
+    "timestamp": 1774933932
+}
+```
 
-### File upload ana ingestion
+#### File upload and ingestion
 * URL: /api/v1/object/upload-ingest
 * Method: POST
 * Content-Type: multipart/form-data
@@ -283,11 +333,11 @@ Response (200 OK):
 }
 ```
 
-### Retrieve and Search
+#### Retrieve and Search
 * URL: /api/v1/object/search
 * Method: POST
 * Content-Type: multipart/form-data
-* Pattern: ASYNC
+* Pattern: SYNC
 * Parameters:
 
 | Field | Type | Required | Description |
@@ -298,7 +348,7 @@ Response (200 OK):
 | filter | object | No | Metadata filters (e.g., {"run_id": "...", "tags": ["class"]}). |
 
 Request:
-```json
+```
 curl --location 'http://127.0.0.1:9011/api/v1/object/search' \
 --header 'Content-Type: application/json' \
 --data '{
@@ -346,7 +396,7 @@ Response (200 OK):
     "timestamp": 1774877744
 }
 ```
-### Resource Download (Video/Image/Document)
+#### Resource Download (Video/Image/Document)
 Download existing resources in Minio.
 
 * URL: /api/v1/object/download/{resource_id}

@@ -6,6 +6,7 @@
 import uuid
 import logging
 from fastapi import UploadFile
+from typing import Optional
 
 logger = logging.getLogger(__name__)
 
@@ -24,10 +25,10 @@ def _try_initialize(self):
             self._error_msg = None
         except (ImportError, ModuleNotFoundError) as e:
             self._error_msg = f"Component missing: {str(e)}"
-            logger.error(f"❌ MinIO component load failed: {self._error_msg}")
+            logger.error(f"MinIO component load failed: {self._error_msg}")
         except Exception as e:
             self._error_msg = f"Initialization failed: {str(e)}"
-            logger.error(f"❌ MinIO connection failed: {self._error_msg}")
+            logger.error(f"MinIO connection failed: {self._error_msg}")
 
     @property
     def is_available(self) -> bool:
@@ -63,7 +64,17 @@ async def get_file_stream(self, file_key: str):
             response = self._store.client.get_object(self._store.bucket, file_key)
             return response
         except Exception as e:
-            logger.error(f"❌ Failed to get file {file_key}: {str(e)}")
+            logger.error(f"Failed to get file {file_key}: {str(e)}")
+            raise e
+
+    async def get_file_content(self, file_key: str, bucket_name: Optional[str] = None) -> bytes:
+        if not self.is_available:
+            raise RuntimeError(f"Storage Service is unavailable: {self._error_msg}")
+        target_bucket = bucket_name or self._store.bucket
+        try:
+            return self._store.get_bytes(file_key)
+        except Exception as e:
+            logger.error(f"Failed to read content for {file_key}: {str(e)}")
             raise e
 
 storage_service = StorageService()