processing: clarify per-file stages and outcomes

Expose extraction, analysis, and replacement progress in the file table; preserve row mapping for duplicates and collapsed archives.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

drhider/scanner.py

@@ -14,7 +14,7 @@ import json
 import random
 import time
 import logging
-from concurrent.futures import ThreadPoolExecutor
+from concurrent.futures import ThreadPoolExecutor, as_completed
 from typing import Dict, List, Optional
 
 from .config import (
@@ -294,13 +294,22 @@ def scan_llm_ner(all_texts: Dict[str, str], mapping: Dict[str, str],
             with ThreadPoolExecutor(max_workers=_LLM_CONCURRENCY) as _ex:
                 futures = [_ex.submit(_call_llm, _build_llm_prompt(c, mapping.keys()), llm_client)
                            for c in chunks]
-                all_ent = []
-                for fut in futures:
+                chunk_results = [None] * len(futures)
+                future_indexes = {future: index for index, future in enumerate(futures)}
+                completed = 0
+                for fut in as_completed(futures):
+                    index = future_indexes[fut]
                     try:
-                        all_ent.extend(fut.result())
+                        chunk_results[index] = fut.result()
                     except Exception as e:
                         llm_failures += 1
                         log.warning("LLM NER chunk failed in %r: %s", fname, e)
+                        chunk_results[index] = []
+                    completed += 1
+                    if file_progress:
+                        file_progress("file_chunk", fname,
+                                      chunks_done=completed, chunks_total=len(chunks))
+                all_ent = [ent for chunk_result in chunk_results for ent in chunk_result]
         else:
             all_ent = []
             for k, c in enumerate(chunks):