Merge pull request #51 from barun-saha/visual

app.py

@@ -20,7 +20,6 @@ from langchain_core.prompts import ChatPromptTemplate
 sys.path.append('..')
 sys.path.append('../..')
 
-import helpers.icons_embeddings as ice
 from global_config import GlobalConfig
 from helpers import llm_helper, pptx_helper, text_helper
 
@@ -56,25 +55,16 @@ def _get_prompt_template(is_refinement: bool) -> str:
 
 
 @st.cache_resource
-def _get_llm():
+def _get_llm(repo_id: str, max_new_tokens: int):
     """
     Get an LLM instance.
 
+    :param repo_id: The model name.
+    :param max_new_tokens: The max new tokens to generate.
     :return: The LLM.
     """
 
-    return llm_helper.get_hf_endpoint()
-
-
-@st.cache_data
-def _get_icons_list() -> List[str]:
-    """
-    Get a list of available icons names without the dir name and file extension.
-
-    :return: A llist of the icons.
-    """
-
-    return ice.get_icons_list()
+    return llm_helper.get_hf_endpoint(repo_id, max_new_tokens)
 
 
 APP_TEXT = _load_strings()
@@ -89,12 +79,19 @@ logger = logging.getLogger(__name__)
 
 texts = list(GlobalConfig.PPTX_TEMPLATE_FILES.keys())
 captions = [GlobalConfig.PPTX_TEMPLATE_FILES[x]['caption'] for x in texts]
-pptx_template = st.sidebar.radio(
-    'Select a presentation template:',
-    texts,
-    captions=captions,
-    horizontal=True
-)
+
+with st.sidebar:
+    pptx_template = st.sidebar.radio(
+        'Select a presentation template:',
+        texts,
+        captions=captions,
+        horizontal=True
+    )
+    st.divider()
+    llm_to_use = st.sidebar.selectbox(
+        'Select an LLM to use:',
+        [f'{k} ({v["description"]})' for k, v in GlobalConfig.HF_MODELS.items()]
+    ).split(' ')[0]
 
 
 def build_ui():
@@ -185,14 +182,12 @@ def set_up_chat_ui():
                 **{
                     'instructions': list_of_msgs,
                     'previous_content': _get_last_response(),
-                    'icons_list': '\n'.join(_get_icons_list())
                 }
             )
         else:
             formatted_template = prompt_template.format(
                 **{
                     'question': prompt,
-                    'icons_list': '\n'.join(_get_icons_list())
                 }
             )
 
@@ -200,12 +195,15 @@ def set_up_chat_ui():
         response = ''
 
         try:
-            for chunk in _get_llm().stream(formatted_template):
+            for chunk in _get_llm(
+                    repo_id=llm_to_use,
+                    max_new_tokens=GlobalConfig.HF_MODELS[llm_to_use]['max_new_tokens']
+            ).stream(formatted_template):
                 response += chunk
 
                 # Update the progress bar
                 progress_percentage = min(
-                    len(response) / GlobalConfig.LLM_MODEL_MAX_OUTPUT_LENGTH, 0.95
+                    len(response) / GlobalConfig.HF_MODELS[llm_to_use]['max_new_tokens'], 0.95
                 )
                 progress_bar.progress(
                     progress_percentage,

global_config.py

@@ -17,10 +17,18 @@ class GlobalConfig:
     A data class holding the configurations.
     """
 
-    HF_LLM_MODEL_NAME = 'mistralai/Mistral-Nemo-Instruct-2407'
+    HF_MODELS = {
+        'mistralai/Mistral-Nemo-Instruct-2407': {
+            'description': 'longer response',
+            'max_new_tokens': 12228
+        },
+        'mistralai/Mistral-7B-Instruct-v0.2': {
+            'description': 'faster, shorter',
+            'max_new_tokens': 8192
+        },
+    }
     LLM_MODEL_TEMPERATURE = 0.2
     LLM_MODEL_MIN_OUTPUT_LENGTH = 100
-    LLM_MODEL_MAX_OUTPUT_LENGTH = 4 * 4096  # tokens
     LLM_MODEL_MAX_INPUT_LENGTH = 400  # characters
 
     HUGGINGFACEHUB_API_TOKEN = os.environ.get('HUGGINGFACEHUB_API_TOKEN', '')

helpers/llm_helper.py

@@ -9,7 +9,6 @@ from langchain_core.language_models import LLM
 from global_config import GlobalConfig
 
 
-HF_API_URL = f"https://api-inference.huggingface.co/models/{GlobalConfig.HF_LLM_MODEL_NAME}"
 HF_API_HEADERS = {"Authorization": f"Bearer {GlobalConfig.HUGGINGFACEHUB_API_TOKEN}"}
 REQUEST_TIMEOUT = 35
 
@@ -28,18 +27,20 @@ http_session.mount('https://', adapter)
 http_session.mount('http://', adapter)
 
 
-def get_hf_endpoint() -> LLM:
+def get_hf_endpoint(repo_id: str, max_new_tokens: int) -> LLM:
     """
     Get an LLM via the HuggingFaceEndpoint of LangChain.
 
-    :return: The LLM.
+    :param repo_id: The model name.
+    :param max_new_tokens: The max new tokens to generate.
+    :return: The HF LLM inference endpoint.
     """
 
-    logger.debug('Getting LLM via HF endpoint')
+    logger.debug('Getting LLM via HF endpoint: %s', repo_id)
 
     return HuggingFaceEndpoint(
-        repo_id=GlobalConfig.HF_LLM_MODEL_NAME,
-        max_new_tokens=GlobalConfig.LLM_MODEL_MAX_OUTPUT_LENGTH,
+        repo_id=repo_id,
+        max_new_tokens=max_new_tokens,
         top_k=40,
         top_p=0.95,
         temperature=GlobalConfig.LLM_MODEL_TEMPERATURE,
@@ -51,69 +52,69 @@ def get_hf_endpoint() -> LLM:
     )
 
 
-def hf_api_query(payload: dict) -> dict:
-    """
-    Invoke HF inference end-point API.
-
-    :param payload: The prompt for the LLM and related parameters.
-    :return: The output from the LLM.
-    """
-
-    try:
-        response = http_session.post(
-            HF_API_URL,
-            headers=HF_API_HEADERS,
-            json=payload,
-            timeout=REQUEST_TIMEOUT
-        )
-        result = response.json()
-    except requests.exceptions.Timeout as te:
-        logger.error('*** Error: hf_api_query timeout! %s', str(te))
-        result = []
-
-    return result
-
-
-def generate_slides_content(topic: str) -> str:
-    """
-    Generate the outline/contents of slides for a presentation on a given topic.
-
-    :param topic: Topic on which slides are to be generated.
-    :return: The content in JSON format.
-    """
-
-    with open(GlobalConfig.SLIDES_TEMPLATE_FILE, 'r', encoding='utf-8') as in_file:
-        template_txt = in_file.read().strip()
-        template_txt = template_txt.replace('<REPLACE_PLACEHOLDER>', topic)
-
-    output = hf_api_query({
-        'inputs': template_txt,
-        'parameters': {
-            'temperature': GlobalConfig.LLM_MODEL_TEMPERATURE,
-            'min_length': GlobalConfig.LLM_MODEL_MIN_OUTPUT_LENGTH,
-            'max_length': GlobalConfig.LLM_MODEL_MAX_OUTPUT_LENGTH,
-            'max_new_tokens': GlobalConfig.LLM_MODEL_MAX_OUTPUT_LENGTH,
-            'num_return_sequences': 1,
-            'return_full_text': False,
-            # "repetition_penalty": 0.0001
-        },
-        'options': {
-            'wait_for_model': True,
-            'use_cache': True
-        }
-    })
-
-    output = output[0]['generated_text'].strip()
-    # output = output[len(template_txt):]
-
-    json_end_idx = output.rfind('```')
-    if json_end_idx != -1:
-        # logging.debug(f'{json_end_idx=}')
-        output = output[:json_end_idx]
-
-    logger.debug('generate_slides_content: output: %s', output)
-
-    return output
+# def hf_api_query(payload: dict) -> dict:
+#     """
+#     Invoke HF inference end-point API.
+#
+#     :param payload: The prompt for the LLM and related parameters.
+#     :return: The output from the LLM.
+#     """
+#
+#     try:
+#         response = http_session.post(
+#             HF_API_URL,
+#             headers=HF_API_HEADERS,
+#             json=payload,
+#             timeout=REQUEST_TIMEOUT
+#         )
+#         result = response.json()
+#     except requests.exceptions.Timeout as te:
+#         logger.error('*** Error: hf_api_query timeout! %s', str(te))
+#         result = []
+#
+#     return result
+
+
+# def generate_slides_content(topic: str) -> str:
+#     """
+#     Generate the outline/contents of slides for a presentation on a given topic.
+#
+#     :param topic: Topic on which slides are to be generated.
+#     :return: The content in JSON format.
+#     """
+#
+#     with open(GlobalConfig.SLIDES_TEMPLATE_FILE, 'r', encoding='utf-8') as in_file:
+#         template_txt = in_file.read().strip()
+#         template_txt = template_txt.replace('<REPLACE_PLACEHOLDER>', topic)
+#
+#     output = hf_api_query({
+#         'inputs': template_txt,
+#         'parameters': {
+#             'temperature': GlobalConfig.LLM_MODEL_TEMPERATURE,
+#             'min_length': GlobalConfig.LLM_MODEL_MIN_OUTPUT_LENGTH,
+#             'max_length': GlobalConfig.LLM_MODEL_MAX_OUTPUT_LENGTH,
+#             'max_new_tokens': GlobalConfig.LLM_MODEL_MAX_OUTPUT_LENGTH,
+#             'num_return_sequences': 1,
+#             'return_full_text': False,
+#             # "repetition_penalty": 0.0001
+#         },
+#         'options': {
+#             'wait_for_model': True,
+#             'use_cache': True
+#         }
+#     })
+#
+#     output = output[0]['generated_text'].strip()
+#     # output = output[len(template_txt):]
+#
+#     json_end_idx = output.rfind('```')
+#     if json_end_idx != -1:
+#         # logging.debug(f'{json_end_idx=}')
+#         output = output[:json_end_idx]
+#
+#     logger.debug('generate_slides_content: output: %s', output)
+#
+#     return output
 
 
 if __name__ == '__main__':

langchain_templates/chat_prompts/initial_template_v4_two_cols_img.txt

@@ -13,12 +13,8 @@ In addition, for each slide, add image keywords based on the content of the resp
 These keywords will be later used to search for images from the Web relevant to the slide content.
 
 In addition, create one slide containing 4 TO 6 icons (pictograms) illustrating some key ideas/aspects/concepts relevant to the topic.
-In this slide, each line of text will begin with the name of a relevant icon enclosed between [[ and ]].
-Select appropriate and exact icon names from the <ICONS> section provided below.
-
-<ICONS>
-{icons_list}
-</ICONS>
+In this slide, each line of text will begin with the name of a relevant icon enclosed between [[ and ]], e.g., [[machine-learning]] and [[fairness]].
+Insert icons only in this slide.
 
 The content of each slide should be VERBOSE, DESCRIPTIVE, and very DETAILED. Each bullet point should be detailed and explanatory, not just short phrases.
 

langchain_templates/chat_prompts/refinement_template_v4_two_cols_img.txt

@@ -14,12 +14,8 @@ In addition, for each slide, add image keywords based on the content of the resp
 These keywords will be later used to search for images from the Web relevant to the slide content.
 
 In addition, create one slide containing 4 TO 6 icons (pictograms) illustrating some key ideas/aspects/concepts relevant to the topic.
-In this slide, each line of text will begin with the name of a relevant icon enclosed between [[ and ]].
-Select appropriate and exact icon names from the <ICONS> section provided below.
-
-<ICONS>
-{icons_list}
-</ICONS>
+In this slide, each line of text will begin with the name of a relevant icon enclosed between [[ and ]], e.g., [[machine-learning]] and [[fairness]].
+Insert icons only in this slide.
 
 The content of each slide should be VERBOSE, DESCRIPTIVE, and very DETAILED. Each bullet point should be detailed and explanatory, not just short phrases.
 

legacy_app.py

@@ -1,294 +0,0 @@
-import pathlib
-import logging
-import tempfile
-from typing import List, Tuple
-
-import json5
-import metaphor_python as metaphor
-import streamlit as st
-
-from helpers import llm_helper, pptx_helper
-from global_config import GlobalConfig
-
-
-APP_TEXT = json5.loads(open(GlobalConfig.APP_STRINGS_FILE, 'r', encoding='utf-8').read())
-GB_CONVERTER = 2 ** 30
-
-
-logger = logging.getLogger(__name__)
-
-
-@st.cache_data
-def get_contents_wrapper(text: str) -> str:
-    """
-    Fetch and cache the slide deck contents on a topic by calling an external API.
-
-    :param text: The presentation topic.
-    :return: The slide deck contents or outline in JSON format.
-    """
-
-    logger.info('LLM call because of cache miss...')
-    return llm_helper.generate_slides_content(text).strip()
-
-
-@st.cache_resource
-def get_metaphor_client_wrapper() -> metaphor.Metaphor:
-    """
-    Create a Metaphor client for semantic Web search.
-
-    :return: Metaphor instance.
-    """
-
-    return metaphor.Metaphor(api_key=GlobalConfig.METAPHOR_API_KEY)
-
-
-@st.cache_data
-def get_web_search_results_wrapper(text: str) -> List[Tuple[str, str]]:
-    """
-    Fetch and cache the Web search results on a given topic.
-
-    :param text: The topic.
-    :return: A list of (title, link) tuples.
-    """
-
-    results = []
-    search_results = get_metaphor_client_wrapper().search(
-        text,
-        use_autoprompt=True,
-        num_results=5
-    )
-
-    for a_result in search_results.results:
-        results.append((a_result.title, a_result.url))
-
-    return results
-
-
-def build_ui():
-    """
-    Display the input elements for content generation. Only covers the first step.
-    """
-
-    # get_disk_used_percentage()
-
-    st.title(APP_TEXT['app_name'])
-    st.subheader(APP_TEXT['caption'])
-    st.markdown(
-        'Powered by'
-        ' [Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2).'
-    )
-    st.markdown(
-        '*If the JSON is generated or parsed incorrectly, try again later by making minor changes'
-        ' to the input text.*'
-    )
-
-    with st.form('my_form'):
-        # Topic input
-        try:
-            with open(GlobalConfig.PRELOAD_DATA_FILE, 'r', encoding='utf-8') as in_file:
-                preload_data = json5.loads(in_file.read())
-        except (FileExistsError, FileNotFoundError):
-            preload_data = {'topic': '', 'audience': ''}
-
-        topic = st.text_area(
-            APP_TEXT['input_labels'][0],
-            value=preload_data['topic']
-        )
-
-        texts = list(GlobalConfig.PPTX_TEMPLATE_FILES.keys())
-        captions = [GlobalConfig.PPTX_TEMPLATE_FILES[x]['caption'] for x in texts]
-
-        pptx_template = st.radio(
-            'Select a presentation template:',
-            texts,
-            captions=captions,
-            horizontal=True
-        )
-
-        st.divider()
-        submit = st.form_submit_button('Generate slide deck')
-
-    if submit:
-        # st.write(f'Clicked {time.time()}')
-        st.session_state.submitted = True
-
-    # https://github.com/streamlit/streamlit/issues/3832#issuecomment-1138994421
-    if 'submitted' in st.session_state:
-        progress_text = 'Generating the slides...give it a moment'
-        progress_bar = st.progress(0, text=progress_text)
-
-        topic_txt = topic.strip()
-        generate_presentation(topic_txt, pptx_template, progress_bar)
-
-    st.divider()
-    st.text(APP_TEXT['tos'])
-    st.text(APP_TEXT['tos2'])
-
-    st.markdown(
-        '![Visitors]'
-        '(https://api.visitorbadge.io/api/visitors?path=https%3A%2F%2Fhuggingface.co%2Fspaces%2Fbarunsaha%2Fslide-deck-ai&countColor=%23263759)'
-    )
-
-
-def generate_presentation(topic: str, pptx_template: str, progress_bar):
-    """
-    Process the inputs to generate the slides.
-
-    :param topic: The presentation topic based on which contents are to be generated.
-    :param pptx_template: The PowerPoint template name to be used.
-    :param progress_bar: Progress bar from the page.
-    """
-
-    topic_length = len(topic)
-    logger.debug('Input length:: topic: %s', topic_length)
-
-    if topic_length >= 10:
-        logger.debug('Topic: %s', topic)
-        target_length = min(topic_length, GlobalConfig.LLM_MODEL_MAX_INPUT_LENGTH)
-
-        try:
-            # Step 1: Generate the contents in JSON format using an LLM
-            json_str = process_slides_contents(topic[:target_length], progress_bar)
-            logger.debug('Truncated topic: %s', topic[:target_length])
-            logger.debug('Length of JSON: %d', len(json_str))
-
-            # Step 2: Generate the slide deck based on the template specified
-            if len(json_str) > 0:
-                st.info(
-                    'Tip: The generated content doesn\'t look so great?'
-                    ' Need alternatives? Just change your description text and try again.',
-                    icon="💡️"
-                )
-            else:
-                st.error(
-                    'Unfortunately, JSON generation failed, so the next steps would lead'
-                    ' to nowhere. Try again or come back later.'
-                )
-                return
-
-            all_headers = generate_slide_deck(json_str, pptx_template, progress_bar)
-
-            # Step 3: Bonus stuff: Web references and AI art
-            show_bonus_stuff(all_headers)
-
-        except ValueError as ve:
-            st.error(f'Unfortunately, an error occurred: {ve}! '
-                     f'Please change the text, try again later, or report it, sharing your inputs.')
-
-    else:
-        st.error('Not enough information provided! Please be little more descriptive :)')
-
-
-def process_slides_contents(text: str, progress_bar: st.progress) -> str:
-    """
-    Convert given text into structured data and display. Update the UI.
-
-    :param text: The topic description for the presentation.
-    :param progress_bar: Progress bar for this step.
-    :return: The contents as a JSON-formatted string.
-    """
-
-    json_str = ''
-
-    try:
-        logger.info('Calling LLM for content generation on the topic: %s', text)
-        json_str = get_contents_wrapper(text)
-    except Exception as ex:
-        st.error(
-            f'An exception occurred while trying to convert to JSON. It could be because of heavy'
-            f' traffic or something else. Try doing it again or try again later.'
-            f'\nError message: {ex}'
-        )
-
-    progress_bar.progress(50, text='Contents generated')
-
-    with st.expander('The generated contents (in JSON format)'):
-        st.code(json_str, language='json')
-
-    return json_str
-
-
-def generate_slide_deck(json_str: str, pptx_template: str, progress_bar) -> List:
-    """
-    Create a slide deck.
-
-    :param json_str: The contents in JSON format.
-    :param pptx_template: The PPTX template name.
-    :param progress_bar: Progress bar.
-    :return: A list of all slide headers and the title.
-    """
-
-    progress_text = 'Creating the slide deck...give it a moment'
-    progress_bar.progress(75, text=progress_text)
-
-    # # Get a unique name for the file to save -- use the session ID
-    # ctx = st_sr.get_script_run_ctx()
-    # session_id = ctx.session_id
-    # timestamp = time.time()
-    # output_file_name = f'{session_id}_{timestamp}.pptx'
-
-    temp = tempfile.NamedTemporaryFile(delete=False, suffix='.pptx')
-    path = pathlib.Path(temp.name)
-
-    logger.info('Creating PPTX file...')
-    all_headers = pptx_helper.generate_powerpoint_presentation(
-        json_str,
-        slides_template=pptx_template,
-        output_file_path=path
-    )
-    progress_bar.progress(100, text='Done!')
-
-    with open(path, 'rb') as f:
-        st.download_button('Download PPTX file', f, file_name='Presentation.pptx')
-
-    if temp:
-        temp.close()
-
-    return all_headers
-
-
-def show_bonus_stuff(ppt_headers: List[str]):
-    """
-    Show bonus stuff for the presentation.
-
-    :param ppt_headers: A list of the slide headings.
-    """
-
-    # Use the presentation title and the slide headers to find relevant info online
-    logger.info('Calling Metaphor search...')
-    ppt_text = ' '.join(ppt_headers)
-    search_results = get_web_search_results_wrapper(ppt_text)
-    md_text_items = []
-
-    for (title, link) in search_results:
-        md_text_items.append(f'[{title}]({link})')
-
-    with st.expander('Related Web references'):
-        st.markdown('\n\n'.join(md_text_items))
-
-    logger.info('Done!')
-
-    # # Avoid image generation. It costs time and an API call, so just limit to the text generation.
-    # with st.expander('AI-generated image on the presentation topic'):
-    #     logger.info('Calling SDXL for image generation...')
-    #     # img_empty.write('')
-    #     # img_text.write(APP_TEXT['image_info'])
-    #     image = get_ai_image_wrapper(ppt_text)
-    #
-    #     if len(image) > 0:
-    #         image = base64.b64decode(image)
-    #         st.image(image, caption=ppt_text)
-    #         st.info('Tip: Right-click on the image to save it.', icon="💡️")
-    #         logger.info('Image added')
-
-
-def main():
-    """
-    Trigger application run.
-    """
-
-    build_ui()
-
-
-if __name__ == '__main__':
-    main()