Merge pull request #429 from christianlouis/copilot/fix-complete-status-error

fix(tests): align "completed" test fixtures with terminal-step guard semantics

app/tasks/process_document.py

@@ -17,6 +17,7 @@ from app.tasks.extract_metadata_with_gpt import extract_metadata_with_gpt
 from app.tasks.process_with_ocr import process_with_ocr
 from app.tasks.retry_config import BaseTaskWithRetry
 from app.utils import get_unique_filepath_with_counter, hash_file, log_task_progress
+from app.utils.step_manager import initialize_file_steps
 from app.utils.text_quality import check_text_quality, detect_pdf_text_source
 
 logger = logging.getLogger(__name__)
@@ -202,6 +203,9 @@ def process_document(
             db.commit()
             db.refresh(new_record)
             logger.info(f"[{task_id}] File record created with ID: {new_record.id}")
+            # Pre-initialize all expected processing steps as "pending" so that
+            # status tracking reflects the complete pipeline from the start.
+            initialize_file_steps(db, new_record.id)
             log_task_progress(
                 task_id,
                 "create_file_record",

app/utils/file_queries.py

@@ -11,6 +11,7 @@ from sqlalchemy import or_
 from sqlalchemy.orm import Query, Session
 
 from app.models import FileProcessingStep, FileRecord
+from app.utils.step_manager import TERMINAL_STEP
 
 
 def apply_status_filter(query: Query, db: Session, status: Optional[str]) -> Query:
@@ -98,6 +99,9 @@ def apply_status_filter(query: Query, db: Session, status: Optional[str]) -> Que
         query = query.filter(FileRecord.id.in_(db.query(subq.c.file_id)))
     elif status == "completed":
         # Files where all real steps are either success or skipped (no failures or in_progress)
+        # and the terminal send_to_all_destinations step has been recorded.
+        # Excluding the terminal-step requirement allows files that only completed the
+        # first few pipeline stages to be falsely labelled as "completed".
         # Exclude duplicates from completed
         query = query.filter(FileRecord.is_duplicate.is_(False))
 
@@ -113,9 +117,19 @@ def apply_status_filter(query: Query, db: Session, status: Optional[str]) -> Que
             .subquery()
         )
 
-        # Select files with real steps that don't have issues
-        query = query.filter(FileRecord.id.in_(db.query(files_with_real_steps.c.file_id))).filter(
-            ~FileRecord.id.in_(db.query(files_with_issues.c.file_id))
+        # Get files that have the terminal processing step recorded
+        files_with_terminal_step = (
+            db.query(FileProcessingStep.file_id)
+            .filter(FileProcessingStep.step_name == TERMINAL_STEP)
+            .distinct()
+            .subquery()
+        )
+
+        # Select files with real steps, no issues, and terminal step present
+        query = (
+            query.filter(FileRecord.id.in_(db.query(files_with_real_steps.c.file_id)))
+            .filter(~FileRecord.id.in_(db.query(files_with_issues.c.file_id)))
+            .filter(FileRecord.id.in_(db.query(files_with_terminal_step.c.file_id)))
         )
     elif status == "duplicate":
         # Files marked as duplicates

app/utils/file_status.py

@@ -7,7 +7,7 @@ from typing import Dict, List
 from sqlalchemy.orm import Session
 
 from app.models import FileProcessingStep, FileRecord, ProcessingLog
-from app.utils.step_manager import get_file_overall_status
+from app.utils.step_manager import TERMINAL_STEP, get_file_overall_status
 
 
 def get_file_processing_status(db: Session, file_id: int) -> Dict:
@@ -150,7 +150,14 @@ def get_files_processing_status(db: Session, file_ids: List[int]) -> Dict[int, D
             elif in_progress_steps > 0:
                 status = "processing"
             elif completed_steps + skipped_steps == total_steps:
-                status = "completed"
+                # Only mark as completed if the terminal processing step has been
+                # recorded. Without this guard, files where later pipeline steps
+                # have not yet started would be falsely marked as "completed".
+                existing_step_names = {s.step_name for s in file_steps}
+                if TERMINAL_STEP in existing_step_names:
+                    status = "completed"
+                else:
+                    status = "pending"
             else:
                 status = "pending"
 

app/utils/step_manager.py

@@ -29,6 +29,10 @@ OPTIONAL_PROCESSING_STEPS = {
     "check_for_duplicates": settings.enable_deduplication,  # Only if deduplication is enabled
 }
 
+# The terminal step is the last mandatory step in the processing pipeline.
+# A file is only considered "completed" once this step has been recorded.
+TERMINAL_STEP = "send_to_all_destinations"
+
 # Combine steps based on configuration
 MAIN_PROCESSING_STEPS = []
 if settings.enable_deduplication:
@@ -252,7 +256,14 @@ def get_file_overall_status(db: Session, file_id: int) -> Dict:
     elif in_progress_steps > 0:
         status = "processing"
     elif completed_steps + skipped_steps == total_steps:
-        status = "completed"
+        # Only mark as completed if the terminal processing step has been recorded.
+        # Without this guard, files where later pipeline steps have not yet started
+        # would be falsely marked as "completed" (e.g. only the first 3 steps ran).
+        existing_step_names = {s.step_name for s in steps}
+        if TERMINAL_STEP in existing_step_names:
+            status = "completed"
+        else:
+            status = "pending"
     else:
         status = "pending"
 
@@ -340,6 +351,17 @@ def get_step_summary(db: Session, file_id: int) -> Dict:
                 main_counts[status] += 1
             main_steps_count += 1
 
+    # Ensure the terminal step is always counted in total_main_steps.
+    # If the terminal step has not been recorded yet, the pipeline is not
+    # complete; counting it as "queued" prevents the status banner from
+    # showing "Completed" before the full pipeline has run.
+    # Note: if TERMINAL_STEP already appears in `steps`, the loop above has
+    # already incremented `main_steps_count` for it, so we only add here when
+    # the step is absent from the DB entirely.
+    if not any(step.step_name == TERMINAL_STEP for step in steps):
+        main_steps_count += 1
+        main_counts["queued"] += 1
+
     return {
         "main": main_counts,
         "uploads": upload_counts,