addressing auto ingest

app.py

@@ -1,4 +1,6 @@
 import os
+import sys
+import subprocess
 
 os.environ.setdefault("LANGCHAIN_TELEMETRY_ENABLED", "false")
 os.environ.setdefault("LANGCHAIN_DISABLE_TELEMETRY", "true")
@@ -7,7 +9,6 @@ os.environ.setdefault("CHROMA_TELEMETRY_ENABLED", "false")
 import streamlit as st
 from src.vectorstore import get_retriever
 from src.qa_chain import make_conversational_chain
-from src.ingest import ingest as run_ingest
 
 st.set_page_config(page_title="Abalone RAG Chatbot", page_icon="🐚")
 
@@ -35,22 +36,6 @@ top_k = st.sidebar.slider(
 data_dir = st.sidebar.text_input("Data directory", value="./data")
 persist_dir = st.sidebar.text_input("Vectorstore directory", value="./vectorstore")
 
-chunk_size = st.sidebar.number_input(
-    "Chunk size",
-    min_value=200,
-    max_value=4000,
-    value=1000,
-    step=100,
-)
-
-chunk_overlap = st.sidebar.number_input(
-    "Chunk overlap",
-    min_value=0,
-    max_value=1000,
-    value=200,
-    step=50,
-)
-
 st.sidebar.markdown("---")
 st.sidebar.caption(
     "If the vectorstore is missing or invalid, the app will attempt to ingest "
@@ -60,32 +45,34 @@ st.sidebar.caption(
 if "chat_history" not in st.session_state:
     st.session_state["chat_history"] = []
 
-if "retriever_initialized" not in st.session_state:
-    st.session_state["retriever_initialized"] = False
-
 def ensure_openai_key() -> bool:
     if not os.environ.get("OPENAI_API_KEY"):
         st.error("OPENAI_API_KEY is not set.")
         return False
     return True
 
+def run_ingest_cli(data_dir: str, persist_dir: str):
+    cmd = [
+        sys.executable,
+        "-m",
+        "src.ingest",
+        "--data-dir",
+        data_dir,
+        "--persist-dir",
+        persist_dir,
+    ]
+    subprocess.run(cmd, check=True)
+
 @st.cache_resource(show_spinner=False)
 def build_or_load_retriever_cached(
         data_dir: str,
         persist_dir: str,
         top_k: int,
-        chunk_size: int,
-        chunk_overlap: int,
 ):
     try:
         return get_retriever(persist_dir=persist_dir, top_k=top_k)
     except Exception:
-        run_ingest(
-            data_dir=data_dir,
-            persist_dir=persist_dir,
-            chunk_size=chunk_size,
-            chunk_overlap=chunk_overlap,
-        )
+        run_ingest_cli(data_dir=data_dir, persist_dir=persist_dir)
         return get_retriever(persist_dir=persist_dir, top_k=top_k)
 
 def get_or_build_retriever_with_ui():
@@ -96,9 +83,13 @@ def get_or_build_retriever_with_ui():
             data_dir=data_dir,
             persist_dir=persist_dir,
             top_k=top_k,
-            chunk_size=chunk_size,
-            chunk_overlap=chunk_overlap,
         )
+    except subprocess.CalledProcessError as e:
+        st.error(
+            "Ingestion failed while building the vectorstore.\n\n"
+            f"Command exit code: {e.returncode}"
+        )
+        return None
     except Exception as e:
         st.error(
             "Could not initialize vectorstore.\n\n"