Update app.py

app.py

@@ -4,7 +4,10 @@ import gradio as gr
 import pandas as pd
 from typing import List, Tuple, Dict, Any
 
-from settings import HEALTHCARE_SETTINGS, GENERAL_CONVERSATION_PROMPT, USE_SCENARIO_ENGINE
+from settings import (
+    HEALTHCARE_SETTINGS, GENERAL_CONVERSATION_PROMPT, USE_SCENARIO_ENGINE,
+    DEBUG_PLAN, COHERE_MODEL_PRIMARY, COHERE_TIMEOUT_S, USE_OPEN_FALLBACKS
+)
 from audit_log import log_event
 from privacy import safety_filter, refusal_reply
 from data_registry import DataRegistry
@@ -13,7 +16,7 @@ from healthcare_analysis import HealthcareAnalyzer
 from scenario_planner import parse_to_plan
 from scenario_engine import ScenarioEngine
 from rag import RAGIndex
-from llm_router import generate_narrative, cohere_chat, open_fallback_chat
+from llm_router import generate_narrative, cohere_chat, open_fallback_chat, _co_client
 
 def _sanitize_text(s: str) -> str:
     if not isinstance(s, str): return s
@@ -35,6 +38,19 @@ def is_healthcare_scenario(text: str, has_files: bool) -> bool:
 def _append_msg(history_messages: List[Dict[str, str]], role: str, content: str) -> List[Dict[str, str]]:
     return (history_messages or []) + [{"role": role, "content": content}]
 
+def ping_cohere():
+    try:
+        cli = _co_client()
+        if not cli:
+            return "Cohere client not initialized. Is COHERE_API_KEY set?"
+        from llm_router import cohere_embed
+        vecs = cohere_embed(["hello", "world"])
+        if vecs and len(vecs) == 2:
+            return "Cohere OK ✅ (embed call succeeded)"
+        return "Cohere reachable, but embedding returned no vectors."
+    except Exception as e:
+        return f"Cohere ping failed: {e}"
+
 def handle(user_msg: str, history_messages: List[Dict[str, str]], files: list) -> Tuple[List[Dict[str, str]], str]:
     try:
         safe_in, blocked_in, reason_in = safety_filter(user_msg, mode="input")
@@ -66,19 +82,16 @@ def handle(user_msg: str, history_messages: List[Dict[str, str]], files: list) -
             datasets = analyzer.comprehensive_analysis(safe_in)
             catalog = _dataset_catalog(datasets)
 
-            # LLM → plan (Cohere API)
             plan = parse_to_plan(safe_in, catalog)
-
-            # Deterministic execution
             structured_md = ScenarioEngine.execute_plan(plan, datasets)
 
-            # Narrative via Cohere API (fallback only if enabled)
             rag_hits = [txt for txt, _ in rag.retrieve(safe_in, k=6)]
             narrative = generate_narrative(safe_in, structured_md, rag_hits)
 
-            reply = _sanitize_text(f"{structured_md}\n\n# Narrative & Recommendations\n\n{narrative}")
+            debug_note = f"\n\n> **Planner note:** {getattr(plan, 'notes', '')}" if (DEBUG_PLAN and getattr(plan, "notes", None)) else ""
+            reply = _sanitize_text(f"{structured_md}\n\n# Narrative & Recommendations\n\n{narrative}{debug_note}")
         else:
-            # General chat via Cohere API
+            # General conversation via Cohere (fallback if enabled)
             prompt = f"{GENERAL_CONVERSATION_PROMPT}\n\nUser: {safe_in}\nAssistant:"
             reply = cohere_chat(prompt) or open_fallback_chat(prompt) or "How can I help further?"
             reply = _sanitize_text(reply)
@@ -97,12 +110,14 @@ def handle(user_msg: str, history_messages: List[Dict[str, str]], files: list) -
 # -------- UI --------
 with gr.Blocks(analytics_enabled=False) as demo:
     gr.Markdown("## Canadian Healthcare AI • Cohere API • Scenario-Agnostic • Deterministic analytics")
-    chat = gr.Chatbot(type="messages", height=520)  # OpenAI-style role/content
-    files = gr.Files(
-        file_count="multiple",
-        type="filepath",
-        file_types=HEALTHCARE_SETTINGS["supported_file_types"]
-    )
+
+    # diagnostics row
+    with gr.Row():
+        ping_btn = gr.Button("Ping Cohere")
+        ping_out = gr.Markdown()
+
+    chat = gr.Chatbot(type="messages", height=520)
+    files = gr.Files(file_count="multiple", type="filepath", file_types=HEALTHCARE_SETTINGS["supported_file_types"])
     msg = gr.Textbox(placeholder="Paste any scenario (Background / Situation / Tasks / Deliverables) or just chat.")
     send = gr.Button("Send")
     clear = gr.Button("Clear")
@@ -111,9 +126,18 @@ with gr.Blocks(analytics_enabled=False) as demo:
         h2, _ = handle(m, h or [], f or [])
         return h2, ""
 
+    ping_btn.click(lambda: ping_cohere(), outputs=[ping_out])
     send.click(_on_send, inputs=[msg, chat, files], outputs=[chat, msg])
     msg.submit(_on_send, inputs=[msg, chat, files], outputs=[chat, msg])
     clear.click(lambda: ([], ""), outputs=[chat, msg])
 
 if __name__ == "__main__":
+    from audit_log import log_event
+    log_event("startup", None, {
+        "cohere_key_present": bool(os.getenv("COHERE_API_KEY")),
+        "cohere_model": COHERE_MODEL_PRIMARY,
+        "timeout_s": COHERE_TIMEOUT_S,
+        "open_fallbacks": USE_OPEN_FALLBACKS
+    })
+    gr.set_static_paths({})
     demo.launch(server_name="0.0.0.0", server_port=int(os.getenv("PORT", "7860")))