Update app.py

app.py

@@ -22,6 +22,8 @@ import openpyxl
 import pptx
 import PyPDF2
 import requests
+from requests.adapters import HTTPAdapter
+from urllib3.util.retry import Retry
 from youtube_transcript_api import YouTubeTranscriptApi
 from youtube_transcript_api._errors import (
     NoTranscriptFound,
@@ -46,9 +48,7 @@ def read_csv(file_path):
         csv_reader = csv.reader(csvfile)
         csv_data = [row for row in csv_reader]
 
-    return ' '.join(
-        [' '.join(row) for row in csv_data]
-    )
+    return ' '.join([' '.join(row) for row in csv_data])
 
 
 def read_text(file_path):
@@ -80,8 +80,7 @@ def read_docx(file_path):
     doc = docx.Document(file_path)
 
     return '\n'.join(
-        paragraph.text
-        for paragraph in doc.paragraphs
+        [paragraph.text for paragraph in doc.paragraphs]
     )
 
 
@@ -105,9 +104,7 @@ def read_xlsx(file_path):
     for row in sheet.iter_rows(values_only=True):
         text_data += (
             ' '.join(
-                str(cell)
-                for cell in row
-                if cell is not None
+                [str(cell) for cell in row if cell is not None]
             )
             + '\n'
         )
@@ -119,10 +116,7 @@ def read_json(file_path):
     with open(file_path, 'r', encoding='utf-8') as f:
         json_data = json.load(f)
 
-    return json.dumps(
-        json_data,
-        ensure_ascii=False
-    )
+    return json.dumps(json_data, ensure_ascii=False)
 
 
 def read_html(file_path):
@@ -154,50 +148,38 @@ def read_xml(file_path):
     )
 
 
-def clean_url_input(url):
+def extract_youtube_video_id(url):
     if not url:
-        return ''
+        return None
 
     url = str(url).strip()
 
     url = url.replace('\\n', '')
     url = url.replace('\n', '')
-    url = url.replace('\r', '')
-
-    markdown_match = re.search(
-        r'\]\((https?://[^)\s]+)\)',
-        url
-    )
-
-    if markdown_match:
-        url = markdown_match.group(1)
+    url = url.replace('\\', '')
 
     markdown_match = re.search(
-        r'\[?(https?://[^\]\s]+)\]?',
-        url
+        r'\]\(\s*(https?://[^)\s]+)',
+        url,
+        flags=re.IGNORECASE
     )
 
     if markdown_match:
         url = markdown_match.group(1)
 
-    url = url.replace('\\', '')
-
-    url = url.strip()
-
-    url = url.rstrip(
-        '.,!?;:)]}'
+    url_match = re.search(
+        r'https?://(?:www\.)?(?:youtube\.com|youtu\.be)[^\s<>"\']+',
+        url,
+        flags=re.IGNORECASE
     )
 
-    return url
-
-
-def extract_youtube_video_id(url):
-    if not url:
-        return None
+    if url_match:
+        url = url_match.group(0)
 
-    url = clean_url_input(url)
+    url = url.rstrip('.,!?;:)]}')
 
     parsed = urlparse(url)
+
     hostname = (
         parsed.hostname or ''
     ).lower()
@@ -212,13 +194,7 @@ def extract_youtube_video_id(url):
             .split('/')[0]
         )
 
-        if re.fullmatch(
-            r'[A-Za-z0-9_-]{11}',
-            video_id
-        ):
-            return video_id
-
-        return None
+        return video_id or None
 
     if hostname in (
         'youtube.com',
@@ -227,43 +203,38 @@ def extract_youtube_video_id(url):
     ):
 
         if parsed.path == '/watch':
+
             video_ids = parse_qs(
                 parsed.query
             ).get('v')
 
-            if not video_ids:
-                return None
-
-            video_id = video_ids[0]
+            if video_ids:
 
-            if re.fullmatch(
-                r'[A-Za-z0-9_-]{11}',
-                video_id
-            ):
-                return video_id
+                video_id = (
+                    video_ids[0]
+                    .split('&')[0]
+                    .strip()
+                )
 
-            return None
+                return video_id or None
 
         for prefix in (
             '/shorts/',
             '/embed/',
             '/live/'
         ):
+
             if parsed.path.startswith(prefix):
+
                 video_id = (
                     parsed.path[
                         len(prefix):
                     ]
                     .split('/')[0]
+                    .strip()
                 )
 
-                if re.fullmatch(
-                    r'[A-Za-z0-9_-]{11}',
-                    video_id
-                ):
-                    return video_id
-
-                return None
+                return video_id or None
 
     return None
 
@@ -275,47 +246,108 @@ def is_youtube_url(url):
     )
 
 
-def fetch_transcript_text(
-    video_id,
-    lang
-):
-    if hasattr(
-        YouTubeTranscriptApi,
-        'get_transcript'
-    ):
-        transcript = (
-            YouTubeTranscriptApi
-            .get_transcript(
-                video_id,
-                languages=[lang]
-            )
+youtube_session = requests.Session()
+
+youtube_session.headers.update({
+    'User-Agent': (
+        'Mozilla/5.0 '
+        '(X11; Linux x86_64) '
+        'AppleWebKit/537.36 '
+        '(KHTML, like Gecko) '
+        'Chrome/131.0.0.0 Safari/537.36'
+    ),
+    'Accept-Language': 'en-US,en;q=0.9',
+    'Accept': (
+        'text/html,application/xhtml+xml,'
+        'application/xml;q=0.9,image/avif,'
+        'image/webp,*/*;q=0.8'
+    ),
+    'Connection': 'close',
+})
+
+retry_strategy = Retry(
+    total=3,
+    connect=3,
+    read=3,
+    backoff_factor=1,
+    status_forcelist=[429, 500, 502, 503, 504],
+    allowed_methods=[
+        'GET',
+        'HEAD'
+    ]
+)
+
+youtube_adapter = HTTPAdapter(
+    max_retries=retry_strategy
+)
+
+youtube_session.mount(
+    'https://',
+    youtube_adapter
+)
+
+youtube_session.mount(
+    'http://',
+    youtube_adapter
+)
+
+
+ytt_api = YouTubeTranscriptApi(
+    http_client=youtube_session
+)
+
+
+def fetch_transcript_text(video_id):
+    try:
+
+        print(
+            'FETCHING YOUTUBE TRANSCRIPT:',
+            repr(video_id)
         )
 
+        transcript = ytt_api.fetch(
+            video_id,
+            languages=['en', 'ar']
+        )
+
+        if hasattr(
+            transcript,
+            'snippets'
+        ):
+
+            return ' '.join(
+                snippet.text
+                for snippet in transcript.snippets
+            )
+
         return ' '.join(
-            entry['text']
-            for entry in transcript
+            snippet.text
+            for snippet in transcript
         )
 
-    ytt_api = YouTubeTranscriptApi()
+    except (
+        NoTranscriptFound,
+        TranscriptsDisabled,
+        VideoUnavailable
+    ):
 
-    fetched = ytt_api.fetch(
-        video_id,
-        languages=[lang]
-    )
+        raise
 
-    return ' '.join(
-        snippet.text
-        for snippet in fetched
-    )
+    except Exception as e:
 
+        print(
+            'YOUTUBE FETCH ERROR:',
+            repr(e)
+        )
 
-def process_youtube_video(
-    url,
-    languages=['en', 'ar']
-):
+        raise
+
+
+def process_youtube_video(url):
     video_id = extract_youtube_video_id(url)
 
     if not video_id:
+
         return (
             'Invalid YouTube video URL. '
             'Please provide a valid YouTube video link.'
@@ -326,67 +358,61 @@ def process_youtube_video(
         repr(video_id)
     )
 
-    transcript_data = []
-    errors = []
+    try:
 
-    for lang in languages:
-        try:
-            transcript = (
-                fetch_transcript_text(
-                    video_id,
-                    lang
-                )
-            )
+        transcript = fetch_transcript_text(
+            video_id
+        )
 
-            if transcript:
-                transcript_data.append(
-                    transcript
-                )
+        if transcript:
 
-        except (
-            NoTranscriptFound,
-            TranscriptsDisabled,
-            VideoUnavailable
-        ) as e:
+            return transcript
 
-            errors.append(
-                f'{lang}: {type(e).__name__}'
-            )
+        return (
+            'The YouTube transcript was retrieved '
+            'but contains no readable text.'
+        )
 
-        except Exception as e:
+    except NoTranscriptFound:
 
-            print(
-                f'YOUTUBE TRANSCRIPT ERROR [{lang}]:',
-                repr(e)
-            )
+        return (
+            'No English or Arabic transcript '
+            'was found for this YouTube video.'
+        )
 
-            errors.append(
-                f'{lang}: {str(e)}'
-            )
+    except TranscriptsDisabled:
 
-    if transcript_data:
-        return ' '.join(
-            transcript_data
+        return (
+            'Transcripts are disabled for this '
+            'YouTube video.'
         )
 
-    print(
-        'YOUTUBE TRANSCRIPT ERRORS:',
-        repr(errors)
-    )
+    except VideoUnavailable:
 
-    return (
-        'Unable to retrieve a transcript for '
-        'this YouTube video. The video may not '
-        'have English or Arabic captions, '
-        'captions may be disabled, or YouTube '
-        'may be blocking transcript access from '
-        'this server.'
-    )
+        return (
+            'The YouTube video is unavailable.'
+        )
+
+    except Exception as e:
+
+        print(
+            'YOUTUBE TRANSCRIPT ERROR:',
+            repr(e)
+        )
+
+        return (
+            'Unable to retrieve the YouTube transcript '
+            'because the server could not establish a '
+            'successful connection with YouTube.\n\n'
+            f'Error: {str(e)}'
+        )
 
 
 def read_web_page(url):
+
     try:
-        url = clean_url_input(url)
+
+        url = str(url).strip()
 
         response = requests.get(
             url,
@@ -404,6 +430,7 @@ def read_web_page(url):
         )
 
         if response.status_code >= 400:
+
             return (
                 'Unable to access webpage. '
                 f'HTTP status: {response.status_code}'
@@ -416,6 +443,7 @@ def read_web_page(url):
         )
 
         if 'text/plain' in content_type:
+
             return response.text
 
         src = response.text
@@ -431,6 +459,7 @@ def read_web_page(url):
             'noscript',
             'svg'
         ]):
+
             element.decompose()
 
         text_data = soup.get_text(
@@ -444,26 +473,27 @@ def read_web_page(url):
         )
 
         if not text_data:
+
             return (
-                'The webpage was accessed '
-                'successfully but contains no '
-                'readable text.'
+                'The webpage was accessed successfully '
+                'but contains no readable text.'
             )
 
         return text_data
 
     except requests.exceptions.Timeout:
+
         return (
             'The webpage request timed out. '
             'Please try another link.'
         )
 
     except requests.exceptions.RequestException as e:
-        return (
-            f'Unable to access webpage: {e}'
-        )
+
+        return f'Unable to access webpage: {e}'
 
     except Exception as e:
+
         return (
             'An error occurred while reading '
             f'the webpage: {e}'
@@ -474,7 +504,9 @@ def read_data(
     file_path_or_url,
     languages=['en', 'ar']
 ):
+
     if not file_path_or_url:
+
         return 'Unsupported type or format.'
 
     file_path_or_url = str(
@@ -484,14 +516,15 @@ def read_data(
     if is_youtube_url(
         file_path_or_url
     ):
+
         return process_youtube_video(
-            file_path_or_url,
-            languages
+            file_path_or_url
         )
 
     if file_path_or_url.startswith(
         ('http://', 'https://')
     ):
+
         return read_web_page(
             file_path_or_url
         )
@@ -501,46 +534,55 @@ def read_data(
     )
 
     if lower_path.endswith('.csv'):
+
         return read_csv(
             file_path_or_url
         )
 
     elif lower_path.endswith('.txt'):
+
         return read_text(
             file_path_or_url
         )
 
     elif lower_path.endswith('.pdf'):
+
         return read_pdf(
             file_path_or_url
         )
 
     elif lower_path.endswith('.docx'):
+
         return read_docx(
             file_path_or_url
         )
 
     elif lower_path.endswith('.pptx'):
+
         return read_pptx(
             file_path_or_url
         )
 
     elif lower_path.endswith('.xlsx'):
+
         return read_xlsx(
             file_path_or_url
         )
 
     elif lower_path.endswith('.json'):
+
         return read_json(
             file_path_or_url
         )
 
     elif lower_path.endswith('.html'):
+
         return read_html(
             file_path_or_url
         )
 
     elif lower_path.endswith('.xml'):
+
         return read_xml(
             file_path_or_url
         )
@@ -549,11 +591,22 @@ def read_data(
 
 
 def normalize_text(text):
-    if not isinstance(text, str):
+
+    if not isinstance(
+        text,
+        str
+    ):
+
         text = str(text)
 
     text = re.sub(
-        r'\*+',
+        r'\\n',
+        ' ',
+        text
+    )
+
+    text = re.sub(
+        r'\\',
         '',
         text
     )
@@ -566,6 +619,7 @@ def normalize_text(text):
     )
 
     for punc in punctuation:
+
         text = text.replace(
             punc,
             ''
@@ -577,7 +631,9 @@ def normalize_text(text):
         text
     )
 
-    words = word_tokenize(text)
+    words = word_tokenize(
+        text
+    )
 
     return ' '.join(words)
 
@@ -593,18 +649,22 @@ llm = HuggingFaceEndpoint(
     do_sample=True,
 )
 
+
 chat_model = ChatHuggingFace(
     llm=llm
 )
 
+
 model_name = (
     'sentence-transformers/all-mpnet-base-v2'
 )
 
+
 embedding_llm = HuggingFaceEmbeddings(
     model_name=model_name
 )
 
+
 db = FAISS.load_local(
     'faiss_index',
     embedding_llm,
@@ -612,7 +672,10 @@ db = FAISS.load_local(
 )
 
 
-def print_like_dislike(x: gr.LikeData):
+def print_like_dislike(
+    x: gr.LikeData
+):
+
     print(
         x.index,
         x.value,
@@ -624,7 +687,9 @@ def user(
     user_message,
     history
 ):
+
     if not len(user_message):
+
         raise gr.Error(
             'Chat messages cannot be empty'
         )
@@ -642,24 +707,29 @@ def user2(
     history,
     link
 ):
-    if not len(user_message) or not len(link):
+
+    if (
+        not len(user_message)
+        or not len(link)
+    ):
+
         raise gr.Error(
             'Chat messages or links cannot be empty'
         )
 
-    link = clean_url_input(link)
+    link = str(link).strip()
     user_message = str(
         user_message
     ).strip()
 
-    if not link:
-        raise gr.Error(
-            'Please provide a valid URL.'
-        )
+    combined_message = (
+        f'URL: {link}\n'
+        f'QUESTION: {user_message}'
+    )
 
     history.append({
         'role': 'user',
-        'content': user_message
+        'content': combined_message
     })
 
     return '', history, link
@@ -670,13 +740,19 @@ def user3(
     history,
     file_path
 ):
-    if not len(user_message) or not file_path:
+
+    if (
+        not len(user_message)
+        or not file_path
+    ):
+
         raise gr.Error(
             'Chat messages or files cannot be empty'
         )
 
     combined_message = (
-        f'{file_path}\n{user_message}'
+        f'{file_path}\n'
+        f'{user_message}'
     )
 
     history.append({
@@ -707,6 +783,7 @@ def Chat_Message(
     history,
     messages1
 ):
+
     user_msg_text = history[-1]['content']
 
     message = HumanMessage(
@@ -717,19 +794,25 @@ def Chat_Message(
         messages1[-1],
         HumanMessage
     ):
+
         messages1 = messages1[:-2]
 
-    messages1.append(message)
+    messages1.append(
+        message
+    )
 
     if len(messages1) >= 8:
+
         messages1 = messages1[-8:]
 
     try:
+
         response = chat_model.invoke(
             messages1
         )
 
     except Exception as e:
+
         error_message = str(e)
 
         print(
@@ -737,18 +820,23 @@ def Chat_Message(
             repr(e)
         )
 
-        start_index = error_message.find(
-            'Input validation error:'
+        start_index = (
+            error_message.find(
+                'Input validation error:'
+            )
         )
 
-        end_index = error_message.find(
-            'and 4096 `max_new_tokens`'
+        end_index = (
+            error_message.find(
+                'and 4096 `max_new_tokens`'
+            )
         )
 
         if (
             start_index != -1
             and end_index != -1
         ):
+
             raise gr.Error(
                 error_message[
                     start_index:end_index
@@ -771,9 +859,12 @@ def Chat_Message(
     })
 
     for character in response.content:
+
         history[-1]['content'] += character
 
-        time.sleep(0.0025)
+        time.sleep(
+            0.0025
+        )
 
         yield history, messages1
 
@@ -782,6 +873,7 @@ def Internet_Search(
     history,
     messages2
 ):
+
     message = str(
         history[-1]['content']
     )
@@ -790,6 +882,7 @@ def Internet_Search(
         messages2[-1],
         HumanMessage
     ):
+
         messages2 = messages2[:-2]
 
     similar_docs = db.similarity_search(
@@ -798,11 +891,16 @@ def Internet_Search(
     )
 
     if similar_docs:
+
         source_knowledge = '\n'.join(
-            x.page_content
-            for x in similar_docs
+            [
+                x.page_content
+                for x in similar_docs
+            ]
         )
+
     else:
+
         source_knowledge = ''
 
     augmented_prompt = f"""
@@ -827,17 +925,22 @@ information from other sources.
         content=augmented_prompt
     )
 
-    messages2.append(msg)
+    messages2.append(
+        msg
+    )
 
     if len(messages2) >= 4:
+
         messages2 = messages2[-4:]
 
     try:
+
         response = chat_model.invoke(
             messages2
         )
 
     except Exception as e:
+
         error_message = str(e)
 
         print(
@@ -845,18 +948,23 @@ information from other sources.
             repr(e)
         )
 
-        start_index = error_message.find(
-            'Input validation error:'
+        start_index = (
+            error_message.find(
+                'Input validation error:'
+            )
         )
 
-        end_index = error_message.find(
-            'and 4096 `max_new_tokens`'
+        end_index = (
+            error_message.find(
+                'and 4096 `max_new_tokens`'
+            )
         )
 
         if (
             start_index != -1
             and end_index != -1
         ):
+
             raise gr.Error(
                 error_message[
                     start_index:end_index
@@ -879,9 +987,12 @@ information from other sources.
     })
 
     for character in response.content:
+
         history[-1]['content'] += character
 
-        time.sleep(0.0025)
+        time.sleep(
+            0.0025
+        )
 
         yield history, messages2
 
@@ -889,6 +1000,7 @@ information from other sources.
 def generate_chart_config(
     description
 ):
+
     system_instructions = """
 You are a Chart.js configuration generator.
 
@@ -980,7 +1092,11 @@ Rules:
     start = raw.find('{')
     end = raw.rfind('}')
 
-    if start == -1 or end == -1:
+    if (
+        start == -1
+        or end == -1
+    ):
+
         raise ValueError(
             'Model did not return a JSON object.\n'
             f'Raw response:\n{raw}'
@@ -995,24 +1111,30 @@ Rules:
         raw
     )
 
-    config = json.loads(raw)
+    config = json.loads(
+        raw
+    )
 
     if 'type' not in config:
+
         raise ValueError(
             "Chart config missing 'type'"
         )
 
     if 'data' not in config:
+
         raise ValueError(
             "Chart config missing 'data'"
         )
 
     if 'labels' not in config['data']:
+
         raise ValueError(
             "Chart config missing 'data.labels'"
         )
 
     if 'datasets' not in config['data']:
+
         raise ValueError(
             "Chart config missing 'data.datasets'"
         )
@@ -1021,6 +1143,7 @@ Rules:
         config['data']['datasets'],
         list
     ):
+
         raise ValueError(
             "'data.datasets' must be a list"
         )
@@ -1032,6 +1155,7 @@ def Chart_Generator(
     history,
     messages3
 ):
+
     message = str(
         history[-1]['content']
     )
@@ -1040,6 +1164,7 @@ def Chart_Generator(
         messages3[-1],
         HumanMessage
     ):
+
         messages3 = messages3[:-2]
 
     if '#chart' in message.lower():
@@ -1052,14 +1177,17 @@ def Chart_Generator(
         )[1].strip()
 
         if not chart_description:
+
             combined_content = (
                 'Please provide chart details after #chart.'
             )
 
         else:
+
             chart_config = None
 
             try:
+
                 chart_config = (
                     generate_chart_config(
                         chart_description
@@ -1075,6 +1203,7 @@ def Chart_Generator(
                 )
 
             except Exception as e:
+
                 print(
                     'CHART GENERATION ERROR:',
                     repr(e)
@@ -1085,10 +1214,10 @@ def Chart_Generator(
                     f'Error: {str(e)}'
                 )
 
-                chart_config = None
-
             if chart_config:
+
                 try:
+
                     config_json = json.dumps(
                         chart_config,
                         separators=(',', ':')
@@ -1122,20 +1251,25 @@ def Chart_Generator(
                         chart_response.status_code
                     )
 
-                    if chart_response.status_code != 200:
+                    if (
+                        chart_response.status_code
+                        != 200
+                    ):
+
                         print(
                             'QUICKCHART RESPONSE:',
                             chart_response.text[:1000]
                         )
 
                         combined_content = (
-                            'QuickChart failed to '
-                            'generate the chart.\n\n'
+                            'QuickChart failed to generate '
+                            'the chart.\n\n'
                             f'HTTP Status: '
                             f'{chart_response.status_code}'
                         )
 
                     else:
+
                         image_html = (
                             f'<img src="{chart_url}" '
                             'alt="Generated Chart" '
@@ -1146,8 +1280,8 @@ def Chart_Generator(
                         )
 
                         chart_summary_prompt = (
-                            'The following Chart.js '
-                            'configuration was generated:\n\n'
+                            'The following Chart.js configuration '
+                            'was generated:\n\n'
                             f'{json.dumps(chart_config, indent=2)}\n\n'
                             'Briefly describe what this chart '
                             'represents.'
@@ -1167,6 +1301,7 @@ def Chart_Generator(
                         ]
 
                         try:
+
                             response = (
                                 chat_model.invoke(
                                     analysis_messages
@@ -1178,6 +1313,7 @@ def Chart_Generator(
                             )
 
                         except Exception as e:
+
                             print(
                                 'CHART ANALYSIS ERROR:',
                                 repr(e)
@@ -1193,6 +1329,7 @@ def Chart_Generator(
                         )
 
                 except Exception as e:
+
                     print(
                         'QUICKCHART REQUEST ERROR:',
                         repr(e)
@@ -1205,21 +1342,27 @@ def Chart_Generator(
                     )
 
     else:
+
         prompt = HumanMessage(
             content=message
         )
 
-        messages3.append(prompt)
+        messages3.append(
+            prompt
+        )
 
         if len(messages3) >= 6:
+
             messages3 = messages3[-6:]
 
         try:
+
             response = chat_model.invoke(
                 messages3
             )
 
         except Exception as e:
+
             error_message = str(e)
 
             print(
@@ -1227,18 +1370,23 @@ def Chart_Generator(
                 repr(e)
             )
 
-            start_index = error_message.find(
-                'Input validation error:'
+            start_index = (
+                error_message.find(
+                    'Input validation error:'
+                )
             )
 
-            end_index = error_message.find(
-                'and 4096 `max_new_tokens`'
+            end_index = (
+                error_message.find(
+                    'and 4096 `max_new_tokens`'
+                )
             )
 
             if (
                 start_index != -1
                 and end_index != -1
             ):
+
                 raise gr.Error(
                     error_message[
                         start_index:end_index
@@ -1255,7 +1403,9 @@ def Chart_Generator(
             )
         )
 
-        combined_content = response.content
+        combined_content = (
+            response.content
+        )
 
     history.append({
         'role': 'assistant',
@@ -1263,27 +1413,72 @@ def Chart_Generator(
     })
 
     for character in combined_content:
+
         history[-1]['content'] += character
 
-        time.sleep(0.0025)
+        time.sleep(
+            0.0025
+        )
 
         yield history, messages3
 
 
 def extract_url_from_text(text):
+
     if not text:
+
         return None
 
-    text = str(text).strip()
+    if isinstance(
+        text,
+        list
+    ):
+
+        parts = []
 
-    text = text.replace(
-        '\\n',
-        ''
-    )
+        for item in text:
+
+            if isinstance(
+                item,
+                dict
+            ):
+
+                value = item.get(
+                    'text',
+                    ''
+                )
+
+                parts.append(
+                    str(value)
+                )
+
+            else:
+
+                parts.append(
+                    str(item)
+                )
+
+        text = ' '.join(parts)
+
+    elif isinstance(
+        text,
+        dict
+    ):
+
+        text = str(
+            text.get(
+                'text',
+                ''
+            )
+        )
+
+    else:
+
+        text = str(text)
 
     text = text.replace(
-        '\n',
-        ''
+        '\\n',
+        ' '
     )
 
     text = text.replace(
@@ -1291,53 +1486,160 @@ def extract_url_from_text(text):
         ''
     )
 
-    markdown_match = re.search(
-        r'\]\((https?://[^)\s]+)\)',
-        text
+    markdown_matches = re.findall(
+        r'\]\(\s*(https?://[^)\s]+)',
+        text,
+        flags=re.IGNORECASE
     )
 
-    if markdown_match:
-        return (
-            markdown_match
-            .group(1)
-            .rstrip('.,!?;:)]}')
+    if markdown_matches:
+
+        return markdown_matches[0].rstrip(
+            '.,!?;:)]}'
         )
 
-    url_match = re.search(
+    url_matches = re.findall(
         r'https?://[^\s<>"\']+',
-        text
+        text,
+        flags=re.IGNORECASE
     )
 
-    if url_match:
-        return (
-            url_match
-            .group(0)
-            .rstrip('.,!?;:)]}')
+    if url_matches:
+
+        url = url_matches[0]
+
+        return url.rstrip(
+            '.,!?;:)]}'
         )
 
     return None
 
 
+def extract_user_query_from_link_message(
+    content,
+    link
+):
+
+    if isinstance(
+        content,
+        list
+    ):
+
+        parts = []
+
+        for item in content:
+
+            if isinstance(
+                item,
+                dict
+            ):
+
+                parts.append(
+                    str(
+                        item.get(
+                            'text',
+                            ''
+                        )
+                    )
+                )
+
+            else:
+
+                parts.append(
+                    str(item)
+                )
+
+        content = ' '.join(parts)
+
+    elif isinstance(
+        content,
+        dict
+    ):
+
+        content = str(
+            content.get(
+                'text',
+                ''
+            )
+        )
+
+    else:
+
+        content = str(content)
+
+    content = content.replace(
+        '\\n',
+        '\n'
+    )
+
+    content = content.replace(
+        '\\',
+        ''
+    )
+
+    content = content.strip()
+
+    question_match = re.search(
+        r'QUESTION\s*:\s*(.*)$',
+        content,
+        flags=re.IGNORECASE | re.DOTALL
+    )
+
+    if question_match:
+
+        return question_match.group(
+            1
+        ).strip()
+
+    if link:
+
+        question = content.replace(
+            link,
+            ''
+        )
+
+        question = re.sub(
+            r'URL\s*:\s*',
+            '',
+            question,
+            flags=re.IGNORECASE
+        )
+
+        return question.strip()
+
+    return content
+
+
 def Link_Scratch(
     history,
-    messages4,
-    link
+    messages4
 ):
-    user_message = str(
-        history[-1]['content']
-    ).strip()
 
-    link = clean_url_input(link)
+    combined_message = (
+        history[-1]['content']
+    )
 
     if isinstance(
         messages4[-1],
         HumanMessage
     ):
+
         messages4 = messages4[:-2]
 
+    link = extract_url_from_text(
+        combined_message
+    )
+
+    user_message = (
+        extract_user_query_from_link_message(
+            combined_message,
+            link
+        )
+    )
+
     print(
-        'USER QUERY:',
-        repr(user_message)
+        'RAW LINK REQUEST:',
+        repr(combined_message)
     )
 
     print(
@@ -1345,14 +1647,23 @@ def Link_Scratch(
         repr(link)
     )
 
+    print(
+        'USER QUERY:',
+        repr(user_message)
+    )
+
     if not link:
+
         response_message = (
-            'Please provide a valid URL starting '
-            'with http:// or https://'
+            'Please provide a valid URL '
+            'starting with http:// or https://'
         )
 
     else:
-        result = read_data(link)
+
+        result = read_data(
+            link
+        )
 
         print(
             'LINK READ RESULT TYPE:',
@@ -1373,7 +1684,10 @@ def Link_Scratch(
         ]
 
         if (
-            isinstance(result, str)
+            isinstance(
+                result,
+                str
+            )
             and (
                 result in error_results
                 or result.startswith(
@@ -1382,22 +1696,38 @@ def Link_Scratch(
                 or result.startswith(
                     'An error occurred while reading'
                 )
+                or result.startswith(
+                    'Unable to retrieve the YouTube transcript'
+                )
+                or result.startswith(
+                    'No English or Arabic transcript'
+                )
+                or result.startswith(
+                    'Transcripts are disabled'
+                )
+                or result.startswith(
+                    'The YouTube video is unavailable'
+                )
             )
         ):
+
             response_message = result
 
         else:
+
             content_data = normalize_text(
                 result
             )
 
             if not content_data:
+
                 response_message = (
                     'The provided link is empty or '
                     'does not contain any meaningful words.'
                 )
 
             else:
+
                 augmented_prompt = f"""
 You are an AI designed to help understand
 and extract information from provided Link Content.
@@ -1410,9 +1740,9 @@ Query:
 Link Content:
 {content_data}
 
-If the query is not related to specific Link Content,
-engage in general conversation or provide relevant
-information from other sources.
+Answer the user's query using the Link Content.
+Do not treat the URL itself or the word "what"
+as part of the URL.
 """
 
                 message = HumanMessage(
@@ -1426,13 +1756,13 @@ information from other sources.
                 messages4 = messages4[-1:]
 
                 try:
-                    response = (
-                        chat_model.invoke(
-                            messages4
-                        )
+
+                    response = chat_model.invoke(
+                        messages4
                     )
 
                 except Exception as e:
+
                     error_message = str(e)
 
                     print(
@@ -1456,6 +1786,7 @@ information from other sources.
                         start_index != -1
                         and end_index != -1
                     ):
+
                         raise gr.Error(
                             error_message[
                                 start_index:end_index
@@ -1482,9 +1813,12 @@ information from other sources.
     })
 
     for character in response_message:
+
         history[-1]['content'] += character
 
-        time.sleep(0.0025)
+        time.sleep(
+            0.0025
+        )
 
         yield history, messages4
 
@@ -1493,6 +1827,7 @@ def insert_line_breaks(
     text,
     every=8
 ):
+
     return '\n'.join(
         text[i:i + every]
         for i in range(
@@ -1503,7 +1838,10 @@ def insert_line_breaks(
     )
 
 
-def display_file_name(file):
+def display_file_name(
+    file
+):
+
     supported_extensions = [
         '.csv',
         '.txt',
@@ -1524,6 +1862,7 @@ def display_file_name(file):
         file_extension.lower()
         in supported_extensions
     ):
+
         file_name = os.path.basename(
             file.name
         )
@@ -1565,6 +1904,7 @@ def File_Interact(
     filepath,
     messages5
 ):
+
     combined_message = str(
         history[-1]['content']
     )
@@ -1573,12 +1913,14 @@ def File_Interact(
         messages5[-1],
         HumanMessage
     ):
+
         messages5 = messages5[:-2]
 
     link = ''
     user_message = ''
 
     if '\n' in combined_message:
+
         link, user_message = (
             combined_message.split(
                 '\n',
@@ -1586,27 +1928,33 @@ def File_Interact(
             )
         )
 
-        user_message = user_message.strip()
+        user_message = (
+            user_message.strip()
+        )
 
     result = read_data(
         filepath
     )
 
     if result == 'Unsupported type or format.':
+
         response_message = result
 
     else:
+
         content_data = normalize_text(
             result
         )
 
         if not content_data:
+
             response_message = (
                 'The file is empty or does not '
                 'contain any meaningful words.'
             )
 
         else:
+
             augmented_prompt = f"""
 You are an AI designed to help understand
 and extract information from provided File Content.
@@ -1636,6 +1984,7 @@ information from other sources.
             messages5 = messages5[-1:]
 
             try:
+
                 response = (
                     chat_model.invoke(
                         messages5
@@ -1643,6 +1992,7 @@ information from other sources.
                 )
 
             except Exception as e:
+
                 error_message = str(e)
 
                 print(
@@ -1666,6 +2016,7 @@ information from other sources.
                     start_index != -1
                     and end_index != -1
                 ):
+
                     raise gr.Error(
                         error_message[
                             start_index:end_index
@@ -1692,9 +2043,12 @@ information from other sources.
     })
 
     for character in response_message:
+
         history[-1]['content'] += character
 
-        time.sleep(0.0025)
+        time.sleep(
+            0.0025
+        )
 
         yield history, messages5
 
@@ -1703,12 +2057,14 @@ def Explore_WebSite(
     history,
     messages6
 ):
+
     message = history[-1]['content']
 
     if isinstance(
         messages6[-1],
         HumanMessage
     ):
+
         messages6 = messages6[:-2]
 
     links = [
@@ -1716,8 +2072,10 @@ def Explore_WebSite(
     ]
 
     result = '\n'.join(
-        read_data(link)
-        for link in links
+        [
+            read_data(link)
+            for link in links
+        ]
     )
 
     content_data = normalize_text(
@@ -1746,17 +2104,22 @@ information from other sources.
         content=augmented_prompt
     )
 
-    messages6.append(msg)
+    messages6.append(
+        msg
+    )
 
     if len(messages6) >= 4:
+
         messages6 = messages6[-4:]
 
     try:
+
         response = chat_model.invoke(
             messages6
         )
 
     except Exception as e:
+
         error_message = str(e)
 
         print(
@@ -1764,18 +2127,23 @@ information from other sources.
             repr(e)
         )
 
-        start_index = error_message.find(
-            'Input validation error:'
+        start_index = (
+            error_message.find(
+                'Input validation error:'
+            )
         )
 
-        end_index = error_message.find(
-            'and 4096 `max_new_tokens`'
+        end_index = (
+            error_message.find(
+                'and 4096 `max_new_tokens`'
+            )
         )
 
         if (
             start_index != -1
             and end_index != -1
         ):
+
             raise gr.Error(
                 error_message[
                     start_index:end_index
@@ -1798,9 +2166,12 @@ information from other sources.
     })
 
     for character in response.content:
+
         history[-1]['content'] += character
 
-        time.sleep(0.0025)
+        time.sleep(
+            0.0025
+        )
 
         yield history, messages6
 
@@ -2094,7 +2465,7 @@ with gr.Blocks() as demo:
             queue=True
         ).then(
             Link_Scratch,
-            [chatbot, messages4, msg1],
+            [chatbot, messages4],
             [chatbot, messages4]
         )
 
@@ -2105,7 +2476,7 @@ with gr.Blocks() as demo:
             queue=True
         ).then(
             Link_Scratch,
-            [chatbot, messages4, msg1],
+            [chatbot, messages4],
             [chatbot, messages4]
         )
 
@@ -2116,7 +2487,7 @@ with gr.Blocks() as demo:
             queue=True
         ).then(
             Link_Scratch,
-            [chatbot, messages4, msg1],
+            [chatbot, messages4],
             [chatbot, messages4]
         )