Update utils/query_parser.py

utils/query_parser.py

@@ -1,178 +1,177 @@
-#!/usr/bin/env python3
-# -*- coding: utf-8 -*-
-
-import re
-from typing import Dict, List, Any
-
-# Import keywords from separate files
-from utils.genres_data import GENRES_KEYWORDS
-from utils.moods_data import MOOD_KEYWORDS
-
-
-def parse_user_query(query: str) -> Dict[str, Any]:
-    """
-    Parses a natural language user query to extract structured tags
-    like genres, moods, target audience, era, decade, specific person, and media type preference.
-
-    Args:
-        query (str): The user's input query string.
-
-    Returns:
-        Dict[str, Any]: A dictionary containing extracted tags.
-            Example: {
-                "genres": ["sci-fi", "thriller"],
-                "mood": ["suspenseful", "dark"],
-                "target_audience": "adult", # or "children", "young_adult"
-                "era": "modern", # or "classic", "contemporary"
-                "decade": "90s", # e.g., "1990s" -> "90s"
-                "specific_person": "Christopher Nolan", # author or director
-                "media_type_preference": "book" # or "movie", or None
-            }
-    """
-    query_lower = query.lower()
-    parsed_tags: Dict[str, Any] = {
-        "genres": [],
-        "mood": [],
-        "target_audience": None,
-        "era": None,
-        "decade": None,
-        "specific_person": None,
-        "media_type_preference": None,
-        "raw_query": query  # Keep original query for debugging/explanation
-    }
-
-    # --- Media Type Preference (strong indicator) ---
-    if re.search(r'\b(movie|film|picture|flick)s?\b', query_lower):
-        parsed_tags["media_type_preference"] = "movie"
-    if re.search(r'\b(book|novel|read|story)s?\b', query_lower):
-        parsed_tags["media_type_preference"] = "book"
-
-    # --- Genres ---
-    for genre, keywords in GENRES_KEYWORDS.items():
-        if any(re.search(r'\b' + re.escape(k) + r'\b', query_lower) for k in keywords):
-            parsed_tags["genres"].append(genre)
-
-    # Remove duplicates and normalize genres (e.g., 'young adult' as genre can be 'target_audience')
-    parsed_tags["genres"] = list(set(parsed_tags["genres"]))
-
-    # --- Moods / Tone ---
-    for mood, keywords in MOOD_KEYWORDS.items():
-        if any(re.search(r'\b' + re.escape(k) + r'\b', query_lower) for k in keywords):
-            if mood not in parsed_tags["mood"]:
-                parsed_tags["mood"].append(mood)
-    parsed_tags["mood"] = list(set(parsed_tags["mood"]))
-
-    # --- Target Audience ---
-    if re.search(r'\b(children|kid|kids|child(?:ren\'s)?|younger audiences?|juvenile)\b', query_lower):
-        parsed_tags["target_audience"] = "children"
-        if "children" in parsed_tags["genres"]: parsed_tags["genres"].remove("children")
-    elif re.search(r'\b(young adult|teen|teens|ya|adolescent)\b', query_lower):
-        parsed_tags["target_audience"] = "young_adult"
-        if "young adult" in parsed_tags["genres"]: parsed_tags["genres"].remove("young adult")
-    elif re.search(r'\b(adult|mature|grown-up|general audiences?)\b', query_lower):
-        parsed_tags["target_audience"] = "adult"
-        if "adult" in parsed_tags["genres"]: parsed_tags["genres"].remove("adult")
-
-    # --- Era ---
-    if re.search(r'\b(classic|classical|old|vintage|timeless)\b', query_lower):
-        parsed_tags["era"] = "classic"
-    elif re.search(r'\b(contemporary|modern|recent|present-day|current)\b', query_lower):
-        parsed_tags["era"] = "contemporary"
-    elif re.search(r'\b(historical|period|past|ancient|medieval|victorian|retro)\b', query_lower):
-        parsed_tags["era"] = "historical"
-    elif re.search(r'\b(future|futuristic)\b', query_lower):
-        parsed_tags["era"] = "future"
-
-
-    # --- Decade ---
-    decade_match = re.search(r'(\d{2}s|(\d{4})s)\b', query_lower)
-    if decade_match:
-        decade_str = decade_match.group(1)
-        if len(decade_str) == 3:  # e.g., '90s'
-            if decade_str.startswith('0'):
-                parsed_tags["decade"] = "2000s"
-            elif decade_str.startswith('10'):
-                parsed_tags["decade"] = "2010s"
-            elif decade_str.startswith('20'):
-                 parsed_tags["decade"] = "2020s"
-            else:
-                parsed_tags["decade"] = f"19{decade_str}"
-        elif len(decade_str) == 5:  # e.g., '1990s'
-            parsed_tags["decade"] = decade_str
-    # Explicitly check for "current decade"
-    if re.search(r'\b(current|recent) decade\b', query_lower) or re.search(r'\b2020s\b', query_lower):
-        parsed_tags["decade"] = "2020s"
-
-
-    # --- Specific Person (Author/Director/Actor) ---
-    person_patterns = [
-        r'\bby\s+([a-zA-Z\s\.]+)\b',
-        r'\b(?:directed\s+by|director)\s+([a-zA-Z\s\.]+)\b',
-        r'\b(?:written\s+by|author)\s+([a-zA-Z\s\.]+)\b',
-        r'\b(?:starring|featuring|with)\s+([a-zA-Z\s\.]+)\b',
-        r'\b(?:from|like)\s+([a-zA-Z\s\.]+)s?\b'
-    ]
-    for pattern in person_patterns:
-        person_match = re.search(pattern, query_lower)
-        if person_match:
-            person_name = person_match.group(1).strip()
-            parsed_tags["specific_person"] = ' '.join([n.capitalize() for n in person_name.split()])
-            break
-
-    # Clean up genres: remove duplicates and ensure audience isn't duplicated
-    parsed_tags["genres"] = list(set(parsed_tags["genres"]))
-    if parsed_tags["target_audience"] == "young_adult" and "young adult" in parsed_tags["genres"]:
-        parsed_tags["genres"].remove("young adult")
-    if parsed_tags["target_audience"] == "children" and "children" in parsed_tags["genres"]:
-        parsed_tags["genres"].remove("children")
-    if parsed_tags["target_audience"] == "adult" and "adult" in parsed_tags["genres"]:
-        parsed_tags["genres"].remove("adult")
-
-    return parsed_tags
-
-
-if __name__ == '__main__':
-    # Test cases for demonstration
-    queries = [
-        "I want a heartwarming drama movie for young adults from the 90s.",
-        "Recommend a thrilling sci-fi book by Isaac Asimov.",
-        "A dark mystery by Agatha Christie.",
-        "Show me action films for kids under 10.",
-        "I need a romantic comedy released in the 2000s.",
-        "Any classic historical fiction?",
-        "looking for something uplifting for ages 18+",
-        "What about a movie directed by Christopher Nolan?",
-        "A book about adventure for children.",
-        "A suspenseful thriller for adults.",
-        "A historical drama set in the 1800s.",
-        "A funny animation from the 80s.",
-        "A contemporary romance novel.",
-        "A classic sci-fi movie directed by Stanley Kubrick.",
-        "I want a thriller by Stephen King.",
-        "A Japanese film like Akira Kurosawa's.",
-        "I'm feeling sad, recommend a melancholic movie.",
-        "Give me an exciting thriller movie.",
-        "I'm in the mood for something lighthearted.",
-        "Looking for a really dark and grim book.",
-        "Need something joyful to watch.",
-        "I need an uplifting and inspiring film.",
-        "Show me a truly gloomy and depressing story.",
-        "Find me a film that's both chaotic and funny.",
-        "I want something thought-provoking and deep.",
-        "Looking for a movie that's really tense and nerve-wracking.",
-        "Something wistful and nostalgic.",
-        "I'm feeling angry, show me something intense and violent.",
-        "Recommend a bizarre and absurd book.",
-        "A beautiful and poignant love story.",
-        "I need a really witty comedy.",
-        "Something raw and gritty.",
-        "A grand, sweeping epic.",
-        "Something that brings tears to my eyes.",
-        "Find me a slow-paced, meditative film.",
-        "A mind-bending psychological thriller."
-    ]
-
-    for q in queries:
-        parsed = parse_user_query(q)
-        print(f"Query: '{q}'")
-        print(f"Parsed: {parsed}\n")
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+
+import re
+from typing import Dict, List, Any
+
+# Import keywords from separate files
+from utils.genres_data import GENRES_KEYWORDS
+from utils.moods_data import MOOD_KEYWORDS
+
+
+def parse_user_query(query: str) -> Dict[str, Any]:
+    """
+    Parses a natural language user query to extract structured tags
+    like genres, moods, target audience, era, decade, specific person, and media type preference.
+
+    Args:
+        query (str): The user's input query string.
+
+    Returns:
+        Dict[str, Any]: A dictionary containing extracted tags.
+            Example: {
+                "genres": ["sci-fi", "thriller"],
+                "mood": ["suspenseful", "dark"],
+                "target_audience": "adult", # or "children", "young_adult"
+                "era": "modern", # or "classic", "contemporary"
+                "decade": "90s", # e.g., "1990s" -> "90s"
+                "specific_person": "Christopher Nolan", # author or director
+                "media_type_preference": "book" # or "movie", or None
+            }
+    """
+    query_lower = query.lower()
+    parsed_tags: Dict[str, Any] = {
+        "genres": [],
+        "mood": [],
+        "target_audience": None,
+        "era": None,
+        "decade": None,
+        "specific_person": None,
+        "media_type_preference": None,
+        "raw_query": query  # Keep original query for debugging/explanation
+    }
+
+    # --- Media Type Preference (strong indicator) ---
+    if re.search(r'\b(movie|film|picture|flick)s?\b', query_lower):
+        parsed_tags["media_type_preference"] = "movie"
+    if re.search(r'\b(book|novel|read|story)s?\b', query_lower):
+        parsed_tags["media_type_preference"] = "book"
+
+    # --- Genres ---
+    for genre, keywords in GENRES_KEYWORDS.items():
+        if any(re.search(r'\b' + re.escape(k) + r'\b', query_lower) for k in keywords):
+            parsed_tags["genres"].append(genre)
+
+    # Remove duplicates and normalize genres (e.g., 'young adult' as genre can be 'target_audience')
+    parsed_tags["genres"] = list(set(parsed_tags["genres"]))
+
+    # --- Moods / Tone ---
+    for mood, keywords in MOOD_KEYWORDS.items():
+        if any(re.search(r'\b' + re.escape(k) + r'\b', query_lower) for k in keywords):
+            if mood not in parsed_tags["mood"]:
+                parsed_tags["mood"].append(mood)
+    parsed_tags["mood"] = list(set(parsed_tags["mood"]))
+
+    # --- Target Audience ---
+    if re.search(r'\b(children|kid|kids|child(?:ren\'s)?|younger audiences?|juvenile)\b', query_lower):
+        parsed_tags["target_audience"] = "children"
+        if "children" in parsed_tags["genres"]: parsed_tags["genres"].remove("children")
+    elif re.search(r'\b(young adult|teen|teens|ya|adolescent)\b', query_lower):
+        parsed_tags["target_audience"] = "young_adult"
+        if "young adult" in parsed_tags["genres"]: parsed_tags["genres"].remove("young adult")
+    elif re.search(r'\b(adult|mature|grown-up|general audiences?)\b', query_lower):
+        parsed_tags["target_audience"] = "adult"
+        if "adult" in parsed_tags["genres"]: parsed_tags["genres"].remove("adult")
+
+    # --- Era ---
+    if re.search(r'\b(classic|classical|old|vintage|timeless)\b', query_lower):
+        parsed_tags["era"] = "classic"
+    elif re.search(r'\b(contemporary|modern|recent|present-day|current)\b', query_lower):
+        parsed_tags["era"] = "contemporary"
+    elif re.search(r'\b(historical|period|past|ancient|medieval|victorian|retro)\b', query_lower):
+        parsed_tags["era"] = "historical"
+    elif re.search(r'\b(future|futuristic)\b', query_lower):
+        parsed_tags["era"] = "future"
+
+
+    # --- Decade ---
+    decade_match = re.search(r'(\d{2}s|(\d{4})s)\b', query_lower)
+    if decade_match:
+        decade_str = decade_match.group(1)
+        if len(decade_str) == 3:  # e.g., '90s'
+            if decade_str.startswith('0'):
+                parsed_tags["decade"] = "2000s"
+            elif decade_str.startswith('10'):
+                parsed_tags["decade"] = "2010s"
+            elif decade_str.startswith('20'):
+                 parsed_tags["decade"] = "2020s"
+            else:
+                parsed_tags["decade"] = f"19{decade_str}"
+        elif len(decade_str) == 5:  # e.g., '1990s'
+            parsed_tags["decade"] = decade_str
+    # Explicitly check for "current decade"
+    if re.search(r'\b(current|recent) decade\b', query_lower) or re.search(r'\b2020s\b', query_lower):
+        parsed_tags["decade"] = "2020s"
+
+
+    # --- Specific Person (Author/Director/Actor) ---
+    person_patterns = [
+        r'\bby\s+([a-zA-Z\s\.]+)\b',
+        r'\b(?:directed\s+by|director)\s+([a-zA-Z\s\.]+)\b',
+        r'\b(?:written\s+by|author)\s+([a-zA-Z\s\.]+)\b',
+        r'\b(?:starring|featuring|with)\s+([a-zA-Z\s\.]+)\b',
+        r'\b(?:from|like)\s+([a-zA-Z\s\.]+)s?\b'
+    ]
+    for pattern in person_patterns:
+        person_match = re.search(pattern, query_lower)
+        if person_match:
+            person_name = person_match.group(1).strip()
+            parsed_tags["specific_person"] = ' '.join([n.capitalize() for n in person_name.split()])
+            break
+
+    # Clean up genres: remove duplicates and ensure audience isn't duplicated
+    parsed_tags["genres"] = list(set(parsed_tags["genres"]))
+    if parsed_tags["target_audience"] == "young_adult" and "young adult" in parsed_tags["genres"]:
+        parsed_tags["genres"].remove("young adult")
+    if parsed_tags["target_audience"] == "children" and "children" in parsed_tags["genres"]:
+        parsed_tags["genres"].remove("children")
+    if parsed_tags["target_audience"] == "adult" and "adult" in parsed_tags["genres"]:
+        parsed_tags["genres"].remove("adult")
+
+    return parsed_tags
+
+
+if __name__ == '__main__':
+    # Test cases for demonstration
+    queries = [
+        "I want a heartwarming drama movie for young adults from the 90s.",
+        "Recommend a thrilling sci-fi book by Isaac Asimov.",
+        "A dark mystery by Agatha Christie.",
+        "Show me action films for kids under 10.",
+        "I need a romantic comedy released in the 2000s.",
+        "Any classic historical fiction?",
+        "looking for something uplifting for ages 18+",
+        "A book about adventure for children.",
+        "A suspenseful thriller for adults.",
+        "A historical drama set in the 1800s.",
+        "A funny animation from the 80s.",
+        "A contemporary romance novel.",
+        "A classic sci-fi movie directed by Stanley Kubrick.",
+        "I want a thriller by Stephen King.",
+        "A Japanese film like Akira Kurosawa's.",
+        "I'm feeling sad, recommend a melancholic movie.",
+        "Give me an exciting thriller movie.",
+        "I'm in the mood for something lighthearted.",
+        "Looking for a really dark and grim book.",
+        "Need something joyful to watch.",
+        "I need an uplifting and inspiring film.",
+        "Show me a truly gloomy and depressing story.",
+        "Find me a film that's both chaotic and funny.",
+        "I want something thought-provoking and deep.",
+        "Looking for a movie that's really tense and nerve-wracking.",
+        "Something wistful and nostalgic.",
+        "I'm feeling angry, show me something intense and violent.",
+        "Recommend a bizarre and absurd book.",
+        "A beautiful and poignant love story.",
+        "I need a really witty comedy.",
+        "Something raw and gritty.",
+        "A grand, sweeping epic.",
+        "Something that brings tears to my eyes.",
+        "Find me a slow-paced, meditative film.",
+        "A mind-bending psychological thriller."
+    ]
+
+    for q in queries:
+        parsed = parse_user_query(q)
+        print(f"Query: '{q}'")
+        print(f"Parsed: {parsed}\n")