fix(control-plane): show pending uploaded documents from server operations (#2420)

Render in-flight and failed file uploads in the Documents view by deriving
them from the server's file_convert_retain operations — no client-side store
or client-generated document ids.

- surface document_id + original_filename on the operations list endpoint
  (already stored in the operation's result_metadata)
- documents-view derives pending/failed rows from those operations, deduped
  against the real document list by document_id, and polls while in-flight
- bridge the brief window where an operation reports completed before the
  document becomes visible in listDocuments, so the row never flickers

Supersedes #2346 (client-side sessionStorage approach). Closes #2314.

Author: nicoloboschi

hindsight-api-slim/hindsight_api/api/http.py

@@ -2596,6 +2596,10 @@ class OperationResponse(BaseModel):
     task_type: str
     items_count: int
     document_id: str | None = None
+    filename: str | None = Field(
+        default=None,
+        description="Original filename for file-conversion operations (file_convert_retain); null for other task types.",
+    )
     created_at: str
     updated_at: str | None = Field(
         default=None,

hindsight-api-slim/hindsight_api/engine/memory_engine.py

@@ -11447,7 +11447,8 @@ async def list_operations(
                         "id": str(row["operation_id"]),
                         "task_type": row["operation_type"],
                         "items_count": result_metadata.get("items_count", 0),
-                        "document_id": None,
+                        "document_id": result_metadata.get("document_id"),
+                        "filename": result_metadata.get("original_filename"),
                         "created_at": row["created_at"].isoformat(),
                         "updated_at": row["updated_at"].isoformat() if row["updated_at"] else None,
                         "status": row["status"],
@@ -12432,6 +12433,7 @@ async def submit_async_file_retain(
                 task_payload=task_payload,
                 result_metadata={
                     "original_filename": file.filename,
+                    "document_id": item["document_id"],
                 },
                 dedupe_by_bank=False,
             )

hindsight-api-slim/tests/test_file_retain.py

@@ -617,6 +617,59 @@ async def read(self):
     assert len(doc["original_text"]) > 0
 
 
+@pytest.mark.asyncio
+async def test_list_operations_surfaces_file_document_id_and_filename(memory_no_llm_verify, sample_txt_content):
+    """list_operations must expose document_id + filename for file_convert_retain ops.
+
+    The control plane derives its pending-upload rows from these fields (it
+    matches an in-flight operation to the real document via document_id and
+    labels the row with the original filename), so both must round-trip from
+    the operation's result_metadata into the list response.
+    """
+    from hindsight_api.models import RequestContext
+
+    bank_id = "test_file_op_fields_bank"
+    context = RequestContext(internal=True)
+    await memory_no_llm_verify.get_bank_profile(bank_id, request_context=context)
+
+    class MockFile:
+        def __init__(self, content, filename, content_type):
+            self.content = content
+            self.filename = filename
+            self.content_type = content_type
+
+        async def read(self):
+            return self.content
+
+    file_items = [
+        {
+            "file": MockFile(sample_txt_content, "report.txt", "text/plain"),
+            "document_id": "doc_op_fields",
+            "context": None,
+            "metadata": {},
+            "tags": [],
+            "timestamp": None,
+            "parser": ["markitdown"],
+        }
+    ]
+
+    await memory_no_llm_verify.submit_async_file_retain(
+        bank_id=bank_id,
+        file_items=file_items,
+        document_tags=None,
+        request_context=context,
+    )
+
+    result = await memory_no_llm_verify.list_operations(
+        bank_id, task_type="file_convert_retain", request_context=context
+    )
+
+    file_ops = [op for op in result["operations"] if op["task_type"] == "file_convert_retain"]
+    assert len(file_ops) == 1
+    assert file_ops[0]["document_id"] == "doc_op_fields"
+    assert file_ops[0]["filename"] == "report.txt"
+
+
 @pytest.mark.asyncio
 async def test_async_file_retain_serializes_datetime_timestamp(memory_no_llm_verify, sample_txt_content):
     """Async file retain should accept Python datetimes in task payloads."""

hindsight-clients/go/api/openapi.yaml

@@ -6982,6 +6982,9 @@ components:
         document_id:
           nullable: true
           type: string
+        filename:
+          nullable: true
+          type: string
         created_at:
           title: Created At
           type: string

hindsight-clients/go/model_operation_response.go

@@ -31,14 +31,15 @@ class OperationResponse(BaseModel):
     task_type: StrictStr
     items_count: StrictInt
     document_id: Optional[StrictStr] = None
+    filename: Optional[StrictStr] = None
     created_at: StrictStr
     updated_at: Optional[StrictStr] = None
     status: StrictStr
     error_message: Optional[StrictStr]
     retry_count: Optional[StrictInt] = None
     next_retry_at: Optional[StrictStr] = None
     progress: Optional[OperationProgress] = None
-    __properties: ClassVar[List[str]] = ["id", "task_type", "items_count", "document_id", "created_at", "updated_at", "status", "error_message", "retry_count", "next_retry_at", "progress"]
+    __properties: ClassVar[List[str]] = ["id", "task_type", "items_count", "document_id", "filename", "created_at", "updated_at", "status", "error_message", "retry_count", "next_retry_at", "progress"]
 
     model_config = ConfigDict(
         populate_by_name=True,
@@ -87,6 +88,11 @@ def to_dict(self) -> Dict[str, Any]:
         if self.document_id is None and "document_id" in self.model_fields_set:
             _dict['document_id'] = None
 
+        # set to None if filename (nullable) is None
+        # and model_fields_set contains the field
+        if self.filename is None and "filename" in self.model_fields_set:
+            _dict['filename'] = None
+
         # set to None if updated_at (nullable) is None
         # and model_fields_set contains the field
         if self.updated_at is None and "updated_at" in self.model_fields_set:
@@ -128,6 +134,7 @@ def from_dict(cls, obj: Optional[Dict[str, Any]]) -> Optional[Self]:
             "task_type": obj.get("task_type"),
             "items_count": obj.get("items_count"),
             "document_id": obj.get("document_id"),
+            "filename": obj.get("filename"),
             "created_at": obj.get("created_at"),
             "updated_at": obj.get("updated_at"),
             "status": obj.get("status"),

hindsight-clients/python/hindsight_client_api/models/operation_response.py

@@ -2755,6 +2755,12 @@ export type OperationResponse = {
    * Document Id
    */
   document_id?: string | null;
+  /**
+   * Filename
+   *
+   * Original filename for file-conversion operations (file_convert_retain); null for other task types.
+   */
+  filename?: string | null;
   /**
    * Created At
    */

hindsight-clients/typescript/generated/types.gen.ts

@@ -376,6 +376,12 @@ function BankSelectorInner() {
       setDocAsync(false);
       setUploadProgress("");
 
+      // Nudge the documents view to surface the new file_convert_retain
+      // operations right away (it derives pending rows from the server).
+      if (typeof window !== "undefined") {
+        window.dispatchEvent(new Event("hindsight:documents-refresh"));
+      }
+
       // Navigate to documents view
       router.push(bankRoute(currentBank!, "?view=documents"));
     } catch {