Update app.py

app.py

@@ -1,12 +1,22 @@
 import streamlit as st
 import json
+import difflib
+import re
 
+# --- Flexible/fuzzy search utilities ---
 COMMON_NAME_KEYS = ["user", "username", "name", "fullName", "firstName", "lastName", "customer"]
 LOGIN_KEYS = ["lastLogin", "login", "loggedIn", "lastLoggedIn", "last_login", "last_logged_in"]
 
 def normalize(s):
-    return str(s).lower().replace("_", " ").replace("-", " ").replace(".", " ").strip()
+    # Lowercase, replace separators with spaces, strip, collapse whitespace
+    return ' '.join(str(s).lower().replace("_", " ").replace("-", " ").replace(".", " ").split())
+
+def is_fuzzy_match(a, b, threshold=0.7):
+    # Both a and b should be normalized strings
+    ratio = difflib.SequenceMatcher(None, a, b).ratio()
+    return ratio >= threshold or a in b or b in a
 
+# --- Fuzzy, nested user search ---
 def recursive_find_user(target_name):
     matches = []
     norm_target = normalize(target_name)
@@ -14,8 +24,8 @@ def recursive_find_user(target_name):
         def _search(obj, path):
             if isinstance(obj, dict):
                 for k, v in obj.items():
-                    # Direct match for string value
-                    if k in COMMON_NAME_KEYS and isinstance(v, str) and norm_target in normalize(v):
+                    # Direct match for string value (fuzzy/partial)
+                    if k in COMMON_NAME_KEYS and isinstance(v, str) and is_fuzzy_match(norm_target, normalize(v)):
                         login_info = {}
                         for lk in LOGIN_KEYS:
                             if lk in obj:
@@ -27,10 +37,10 @@ def recursive_find_user(target_name):
                             "file": file_name,
                             "login_info": login_info
                         })
-                    # Nested dict, e.g., user: { username: ... }
+                    # Nested dict (fuzzy/partial)
                     if k in COMMON_NAME_KEYS and isinstance(v, dict):
                         for nk, nv in v.items():
-                            if nk in COMMON_NAME_KEYS and norm_target in normalize(nv):
+                            if nk in COMMON_NAME_KEYS and is_fuzzy_match(norm_target, normalize(nv)):
                                 login_info = {}
                                 for lk in LOGIN_KEYS:
                                     if lk in obj:
@@ -49,6 +59,7 @@ def recursive_find_user(target_name):
         _search(data, [])
     return matches
 
+# --- Show all user keys/values (for debug) ---
 def show_all_users():
     found = []
     for file_name, data in st.session_state.json_data.items():
@@ -69,13 +80,13 @@ def show_all_users():
         recursive(data, [])
     return found
 
-import re
+# --- User query handler ---
 def handle_user_query(query):
     patterns = [
-        r"(?:last\s*login.*?for|when\s+did)\s+([a-zA-Z0-9 _\-\.]+)",
-        r"when\s+was\s+([a-zA-Z0-9 _\-\.]+)\s+last\s+(?:login|logged\s*in)",
-        r"last\s*login\s*of\s+([a-zA-Z0-9 _\-\.]+)",
-        r"(?:info|details|record) for\s+([a-zA-Z0-9 _\-\.]+)"
+        r"(?:last\s*login.*?for|when\s+did)\s+([a-zA-Z0-9 _\-\.@]+)",
+        r"when\s+was\s+([a-zA-Z0-9 _\-\.@]+)\s+last\s+(?:login|logged\s*in)",
+        r"last\s*login\s*of\s+([a-zA-Z0-9 _\-\.@]+)",
+        r"(?:info|details|record) for\s+([a-zA-Z0-9 _\-\.@]+)"
     ]
     found_name = None
     for pat in patterns:
@@ -84,7 +95,8 @@ def handle_user_query(query):
             found_name = m.group(1).strip()
             break
     if not found_name:
-        m = re.search(r"([A-Z][a-z]+(?: [A-Z][a-z]+)*)", query)
+        # Fallback: look for any word with at least 3 letters (handles very short names too)
+        m = re.search(r"([A-Za-z0-9][A-Za-z0-9 _\-\.@]{2,})", query)
         if m:
             found_name = m.group(1).strip()
     if found_name:
@@ -101,7 +113,7 @@ def handle_user_query(query):
     else:
         return "Sorry, I can only answer direct user info queries (e.g., 'When did Bob the Builder last login?')."
 
-# --- Streamlit UI ---
+# --- Streamlit UI setup ---
 if "json_data" not in st.session_state:
     st.session_state.json_data = {}
 if "messages" not in st.session_state:
@@ -111,8 +123,8 @@ if "temp_input" not in st.session_state:
 if "files_loaded" not in st.session_state:
     st.session_state.files_loaded = False
 
-st.set_page_config(page_title="Instant JSON Q&A", layout="wide")
-st.title("Instant JSON-Backed AI Q&A (Now Handles Nested Users!)")
+st.set_page_config(page_title="Flexible JSON User Q&A", layout="wide")
+st.title("Instant JSON-Backed AI Q&A (Fuzzy Flexible Search!)")
 
 uploaded_files = st.sidebar.file_uploader(
     "Choose one or more JSON files", type="json", accept_multiple_files=True
@@ -132,7 +144,7 @@ elif not uploaded_files:
     st.session_state.json_data.clear()
     st.session_state.files_loaded = False
 
-st.markdown("### Ask about any user directly (e.g. 'When did Bob the Builder last login?')")
+st.markdown("### Ask about any user (partial/typo/fuzzy OK) — try: `bob`, `bob the builder`, `astrofan`, `alice`, `johnny.appleseed`")
 for msg in st.session_state.messages:
     if msg["role"] == "user":
         st.markdown(f"<div style='color: #4F8BF9;'><b>User:</b> {msg['content']}</div>", unsafe_allow_html=True)