Update app.py

app.py

@@ -2945,7 +2945,7 @@ def iterative_deep_research_gen(initial_query: str, reportstyle: str, breadth: i
                     "title": title,
                     "summary": analysis_summary,
                     "clean_content": clean_content,
-                    "raw_excerpt": raw_content[:200]
+                    "raw_excerpt": raw_content[:1000]
                 })
                 
                 if analysis.get("relevant", "no").lower() == "yes":
@@ -2973,11 +2973,11 @@ def iterative_deep_research_gen(initial_query: str, reportstyle: str, breadth: i
             process_log += "Appended additional clarifications to the context.\n"
         progress_pct = int((iteration / depth) * 100)
         yield (f"Progress: {progress_pct}%", None, None, None, None)
-    
-    # chunk and filter all crumbs if breadth>3 and depth>2
-    filtered_crumbs_list = crumbs_list
+
     if breadth > 3 and depth > 2:
         filtered_crumbs_list = filter_crumbs_in_batches(crumbs_list, initial_query, followup_clarifications)
+    else:
+        filtered_crumbs_list = crumbs_list
 
     # Now build aggregated crumb text from filtered_crumbs_list only
     aggregated_crumbs = "\n\n".join([
@@ -3000,7 +3000,7 @@ def iterative_deep_research_gen(initial_query: str, reportstyle: str, breadth: i
     # We convert processed_queries to a string suitable for storing
     all_processed_queries_str = "\n".join(sorted(processed_queries))
 
-    yield ("", final_report, process_log, crumbs_list, all_processed_queries_str)
+    yield ("", final_report, process_log, aggregated_crumbs, all_processed_queries_str)
 
 def filter_crumbs_in_batches(crumbs_list: list, initial_query: str, clarifications: str) -> list:
     accepted = []
@@ -3020,11 +3020,11 @@ And the provided clarifications:
 
 Previously selected batch:
 
-Now provide the JSCON
+Now provide the JSON
 """
         listing = []
         for idx, c in enumerate(batch):
-            snippet_for_prompt = c["summary"][:500]  # short snippet
+            snippet_for_prompt = c["summary"][:1000]  # short snippet
             listing.append(f"Crumb {idx}: {snippet_for_prompt}")
         prompt += "\n".join(listing)
 
@@ -3035,6 +3035,8 @@ Return a JSON object with structure:
  "1": "yes" or "no",
  ...
 }
+
+No code fences, no 'json' mentioned before the result, only the json result formatted.
 """
         decision_str = openai_call(prompt, model="o3-mini", temperature=0, max_tokens_param=1500)
         # parse JSON