all of the things

README.md

@@ -4,7 +4,7 @@ emoji: 🛠️
 colorFrom: red
 colorTo: red
 sdk: gradio
-sdk_version: 3.40.1
+sdk_version: 4.37.2
 app_file: app.py
 pinned: false
 ---

app.py

@@ -2,29 +2,30 @@ import gradio as gr
 import pandas as pd
 import openai
 from ui.app_theme import SoftBlue
-from image_gen import ui_image_generation
-from summarization import ui_summarize
 from utilities import constants,api_keys
-from recording_analysis import ui_recording_analysis 
 from prompts import ui_prompt_builder
 from chat_bot import ui_simple_chat
+''' Disabled 
+from image_gen import ui_image_generation
+from summarization import ui_summarize
+from recording_analysis import ui_recording_analysis 
 from slack_sentiment_analysis import ui_sentiment_analysis
 from image_to_text import image_to_text_translator as ui_image_to_text
-from text_analysis import ui_text_analysis
+from text_analysis import ui_text_analysis'''
 
 openai.api_key = api_keys.APIKeys().get_key('OPENAI_API_KEY')
 
 prompt_builder=ui_prompt_builder.ui
+usaa_advisor=ui_simple_chat.ui
+''' Disabled
 summarize=ui_summarize.ui
 image_generator=ui_image_generation.ui
 video_analysis=ui_recording_analysis.ui
 slack_sentiment=ui_sentiment_analysis.ui
-usaa_advisor=ui_simple_chat.ui
-image_to_text=ui_image_to_text.ui
-text_analysis=ui_text_analysis.ui
+#git_advisor=ui_simple_chat_git.ui
+#image_to_text=ui_image_to_text.ui
+#text_analysis=ui_text_analysis.ui'''
 
-ui = gr.TabbedInterface([prompt_builder, text_analysis, image_generator, video_analysis,slack_sentiment, usaa_advisor, summarize], 
-                        (constants.UI_1, constants.UI_2, constants.UI_3,constants.UI_4, constants.UI_5,constants.UI_7, constants.UI_8), 
-                        theme=SoftBlue())
+ui = gr.TabbedInterface([prompt_builder, usaa_advisor], (constants.UI_1, "Policy Chat"),theme=SoftBlue())
 
 ui.launch()

chat_bot/simple_chat.py

@@ -1,12 +1,158 @@
-import os, openai,string,csv
+import os
+import openai
+import string
+import csv
+import re
 import pandas as pd
 from utilities import constants, api_keys, clean_text, prompt_constants
 from translators import translate_pdf_to_text
 from embedding_tools import create_embedding
+from chat_bot import process_policies
 
 openai.api_key = api_keys.APIKeys().get_key('OPENAI_API_KEY')
 messages=[]
 
+
+
+def flatten_json(json_obj, prefix=''):
+    items = []
+
+    def flatten(json_fragment, current_prefix):
+        if isinstance(json_fragment, dict):
+            for key, value in json_fragment.items():
+                new_key = f'{current_prefix}{key}_' if current_prefix else key
+                if isinstance(value, (dict, list)):
+                    flatten(value, new_key)
+                else:
+                    # Remove newlines from value
+                    cleaned_value = str(value).replace('\n', ' ')
+                    items.append((new_key, cleaned_value))
+        elif isinstance(json_fragment, list):
+            for index, item in enumerate(json_fragment):
+                new_key = f'{current_prefix}{index}_'
+                if isinstance(item, (dict, list)):
+                    flatten(item, new_key)
+                else:
+                    # Remove newlines from value
+                    cleaned_value = str(item).replace('\n', ' ')
+                    items.append((new_key, cleaned_value))
+
+    flatten(json_obj, prefix)
+    return items
+
+def camel_case_to_readable(text):
+    text = re.sub(r'(?<!^)(?=[A-Z])', ' ', text)
+    text = text.replace('_', ' ')
+    text = text.lower()
+    text = text.capitalize()
+    return text
+
+def make_readable_key(key):
+    key = key.replace('_', ' ')
+    key = re.sub(r'\b0_', '(1) ', key)
+    key = re.sub(r'\b1_', '(2) ', key)
+    key = re.sub(r'\b2_', '(3) ', key)
+    key = ' '.join([word.capitalize() for word in key.split()])
+    key = key.replace('Is', 'is').replace('Has a', 'has a')
+    return key
+
+def process_list(items):
+    readable_items = []
+    for item in items:
+        key, value = item.split(': ', 1)
+        readable_key = make_readable_key(key)
+        readable_value = value.replace('\n', ' ')
+        readable_items.append(f'{readable_key}: {readable_value}')
+    return readable_items
+
+def convert_flattened_to_readable(flattened_items):
+    readable_items = []
+    for key, value in flattened_items:
+        readable_key = make_readable_key(key)
+        readable_items.append(f'{readable_key}: {value}')
+    return readable_items
+
+def convert_json_to_single_column_csv(json_obj):
+    flattened_items = flatten_json(json_obj)
+    readable_items = convert_flattened_to_readable(flattened_items)
+    flat_text = '\n'.join(readable_items)
+    return flat_text
+
+
+def flatten_json_to_single_column(json_obj, prefix=''):
+    print("in flatten_json_to_single_column, json_obj:", json_obj, "prefix:", prefix)
+
+    items = []
+    if isinstance(json_obj, dict):
+        for key, value in json_obj.items():
+            if isinstance(value, dict):
+                items.extend(flatten_json_to_single_column(value, f'{prefix}{key}_'))
+            elif isinstance(value, list):
+                for index, item in enumerate(value):
+                    items.extend(flatten_json_to_single_column(item, f'{prefix}{key}_{index}_'))
+            else:
+                items.append(f'{prefix}{key}: {value}')
+    elif isinstance(json_obj, list):
+        for index, item in enumerate(json_obj):
+            if isinstance(item, dict):
+                items.extend(flatten_json_to_single_column(item, f'{prefix}{index}_'))
+            else:
+                items.append(f'{prefix}{index}: {item}')
+    else:
+        items.append(f'{prefix[:-1]}: {json_obj}')  # Remove trailing underscore
+    return items
+
+def convert_json_to_single_column_csv_old(json_obj):
+    flat_data = flatten_json_to_single_column(json_obj)
+    flat_text = '\n'.join(flat_data)
+    return flat_text
+
+def write_flat_text_to_csv(flat_text, csv_path):
+    # Open a CSV file to write the data
+    with open(csv_path, 'w', newline='') as csvfile:
+        csv_writer = csv.writer(csvfile)
+        
+        # Write each line as a row in the CSV file
+        for line in flat_text:
+            csv_writer.writerow([line])
+
+
+## entry point edited, input=pdf, output=csv
+def CreateEmbeddings(policy_input, policy_output):
+    print("in createembeddings, policy_input:", policy_input, "policy_output:", policy_output)
+    content=process_policies.convert_pdf_to_text_ocr(policy_input, constants.POLICY_TXT_PATH)
+
+    print("\n***content after extract policy:", content)
+    df=None
+    if content:
+        print("**content before:", content)
+        schema=process_policies.create_schema(content)
+        print ("**schema:", schema)
+        flat_txt = flatten_json_to_single_column(schema)
+        readable= process_list(flat_txt)
+        for item in readable:
+            print(item)
+        write_flat_text_to_csv(readable,policy_output)
+        print ("csv writtens")
+        create_embedding.CreateEmbeddingsFlatPolicy(policy_output, constants.POLICY_PKL_PATH)
+    print("df:", df)
+    return df
+
+def CreateEmbeddingsOriginal(input_path, output_path):
+    print("in createembeddingsoriginal, input_path:", input_path, "output_path:", output_path)
+    if os.path.exists(constants.PKL_PATH):
+        df = pd.read_pickle(constants.PKL_PATH)
+        create_embedding.CreateEmbeddingsFlat(constants.CSV_PATH, constants.PKL_PATH)
+        return df
+    
+    transcript= translate_pdf_to_text.TranslatePdfToText(input_path, output_path)
+    text_chunks=CreateChunks(transcript, constants.EMBEDDING_CHUNK_LENGTH)
+    CreateCSV(text_chunks)
+    create_embedding.CreateEmbeddingsFlat(constants.CSV_PATH, constants.PKL_PATH)
+    df = pd.read_pickle(constants.PKL_PATH)
+    return df
+
+
 def CreateCSV(text_chunks):
     with open(constants.CSV_PATH, 'w', newline='') as csvfile:
         csv_writer = csv.writer(csvfile)
@@ -36,19 +182,6 @@ def CreateChunks(transcript, length):
         text_chunks.append(transcript[start_idx:])
     return text_chunks
 
-def CreateEmbeddings(input_path, output_path):
-    if os.path.exists(constants.PKL_PATH):
-        df = pd.read_pickle(constants.PKL_PATH)
-        create_embedding.CreateEmbeddingsFlat(constants.CSV_PATH, constants.PKL_PATH)
-        return df
-    
-    transcript= translate_pdf_to_text.TranslatePdfToText(input_path, output_path)
-    text_chunks=CreateChunks(transcript, constants.EMBEDDING_CHUNK_LENGTH)
-    CreateCSV(text_chunks)
-    create_embedding.CreateEmbeddingsFlat(constants.CSV_PATH, constants.PKL_PATH)
-    df = pd.read_pickle(constants.PKL_PATH)
-    return df
-
 def Completion(messages):
     response = openai.ChatCompletion.create(
         model="gpt-3.5-turbo", 
@@ -58,6 +191,7 @@ def Completion(messages):
     return cleaned_text
 
 def QueryEmbeddingsSimple(query):
+    print ("in QueryEmbeddingsSimple")
     global messages
     best_answer= create_embedding.QueryEmbeddingsFlat(query)
     prompt= prompt_constants.USAA_AR_EXPERT_PROMPT

chat_bot/ui_chat_bot.py

@@ -66,7 +66,7 @@ with gr.Blocks() as ui:
                     sentiment=gr.Checkbox(label=constants.SENTIMENT, value=False, info=constants.SENTIMENT_INFO)
                     emotion=gr.Checkbox(label=constants.EMOTION, value=False, info=constants.EMOTION_INFO)
             gr.Markdown(constants.MICROPHONE_MD)
-            audio = gr.Audio(label=constants.ASK_A_QUESTION, source="microphone", type="filepath")
+            audio = gr.Audio(label=constants.ASK_A_QUESTION)
             with gr.Row():
                 clear=gr.Button(value=constants.CLEAR)
                 submit=gr.Button(value=constants.SUBMIT, variant="primary")

chat_bot/ui_simple_chat.py

@@ -7,9 +7,9 @@ import random
 import time
 
 def InitDF():
-    global commDF
-    commDF=pd.DataFrame({"role": [""], "content": [""] })
-    simple_chat.CreateEmbeddings(constants.PDF_INPUT_PATH, constants.TXT_OUTPUT_PATH)
+    global gitDF
+    gitDF=pd.DataFrame({"role": [""], "content": [""] })
+    simple_chat.CreateEmbeddings(constants.POLICY_INPUT, constants.POLICY_OUTPUT)
 
 def Respond(message, chat_history):
     bot_message,df=simple_chat.QueryEmbeddingsSimple(message)
@@ -28,8 +28,8 @@ with gr.Blocks() as ui:
     question= gr.Textbox (label=constants.QUESTIONS_PREFIX, value=constants.QUESTIONS_AR_EXPERT)
     chatbot = gr.Chatbot(label=constants.CHAT_BOT, height=constants.CHAT_BOT_HEIGHT)
     msg = gr.Textbox(label=constants.CHAT_BOT_INPUT)    
-    commDF = gr.DataFrame(type="pandas", value=pd.DataFrame({"role": [""], "content": [""] }), wrap=True, label=constants.OPENAI_LOG)
+    gitDF = gr.DataFrame(type="pandas", value=pd.DataFrame({"role": [""], "content": [""] }), wrap=True, label=constants.OPENAI_LOG)
     clear = gr.ClearButton([msg, chatbot])
-    msg.submit(Respond, [msg, chatbot], [msg, chatbot,commDF])
+    msg.submit(Respond, [msg, chatbot], [msg, chatbot,gitDF])
 
 InitDF()

embedding_tools/create_embedding.py

@@ -14,6 +14,18 @@ def CreateEmbeddingsQA(input_path, output_path):
         benefit_info['embedding'] = benefit_info['answer'].apply(lambda row: get_embedding(str(row), engine='text-embedding-ada-002'))
         benefit_info.to_pickle(output_path) 
 
+def CreateEmbeddingsFlatPolicy(input_path, output_path):
+    global embedding_info
+    print(" CreateEmbeddingsFlat input_path: "+input_path + " output_path: "+output_path)
+    embedding_info = pd.read_csv(input_path)
+    embedding_info.columns = ['data']
+    embedding_info['embedding'] = embedding_info['data'].apply(lambda row: get_embedding(str(row), engine='text-embedding-ada-002'))
+    embedding_info.to_pickle(output_path) 
+    print("Loaded from CSV file.")
+
+    print(f"embedding_info type: {type(embedding_info)}")
+    print(f"embedding_info is None: {embedding_info is None}")
+
 def CreateEmbeddingsFlat(input_path, output_path):
     global embedding_info
     print(" CreateEmbeddingsFlat input_path: "+input_path + " output_path: "+output_path)

embedding_tools/embeddings.py

@@ -2,6 +2,7 @@ from utilities import constants
 import embedding_tools.create_embedding as create_embedding
 
 def ApplyEmbeddings(message, persona):    
+    print("persona: ", persona)
     if(persona==constants.PERSONA_HR_EXPERT):
         return create_embedding.QueryEmbeddings(message)
     else:

image_to_text/image_to_text_translator.py

@@ -53,7 +53,7 @@ with gr.Blocks() as ui:
             gr.Markdown(constants.DIRECTIONS_MD)
             gr.Markdown(constants.AR_ADVISOR_DIRECTIONS)
     with gr.Row():
-        imageBlock = gr.Image(label=constants.IMAGE_INPUT, source="upload")
+        imageBlock = gr.Image(label=constants.IMAGE_INPUT)
         summaryBlock= gr.Textbox (label=constants.SUMMARY)
     with gr.Row():
         submitBtn=gr.Button(value=constants.EXTRACT_SUMMARY, variant="primary")

image_to_text/ui_image_to_text.py

@@ -36,7 +36,7 @@ with gr.Blocks() as ui:
             gr.Markdown(value=constants.RECORDING_ANALYSIS_DIRECTIONS)
 
     with gr.Row():
-        videoBlock = gr.Video(label=constants.VIDEO_INPUT, source="upload")
+        videoBlock = gr.Video(label=constants.VIDEO_INPUT)
         summaryBlock= gr.Textbox (label=constants.SUMMARY)
     with gr.Column():
         gr.Markdown(constants.TOPICS_MD)

prompts/ui_prompt_builder.py

@@ -79,7 +79,7 @@ with gr.Blocks() as ui:
                               info=constants.VOICE_INFO)
             
             gr.Markdown(constants.MICROPHONE_MD)
-            audio = gr.Audio(label=constants.ASK_A_QUESTION, source="microphone", type="filepath")
+            audio = gr.Audio(label=constants.ASK_A_QUESTION,  type="filepath")
             with gr.Row():
                 clear=gr.Button(value=constants.CLEAR)
                 submit=gr.Button(value=constants.SUBMIT, variant="primary")

recording_analysis/ui_recording_analysis.py

@@ -25,6 +25,23 @@ def ClearScreen():
     videoBlock=None
     return "","",pd.DataFrame({"role": [""], "content": [""] })
 
+'''Alternate way to manage source
+def process_video(video):
+    # Here you can add any processing you might want to perform on the video
+    return video
+
+with gr.Blocks() as demo:
+    video_input = gr.Video(label="Upload your video")
+    submit_button = gr.Button("Submit")
+    video_output = gr.Video(label="Processed Video")
+
+    submit_button.click(
+        fn=process_video,
+        inputs=video_input,
+        outputs=video_output
+    )
+'''
+
 with gr.Blocks() as ui:
     label2 = gr.Label(show_label=False, value=constants.RECORDING_ANALYSIS, container=False)
     with gr.Row():
@@ -36,7 +53,7 @@ with gr.Blocks() as ui:
             gr.Markdown(value=constants.RECORDING_ANALYSIS_DIRECTIONS)
 
     with gr.Row():
-        videoBlock = gr.Video(label=constants.VIDEO_INPUT, source="upload")
+        videoBlock = gr.Video(label=constants.VIDEO_INPUT)
         summaryBlock= gr.Textbox (label=constants.SUMMARY)
     with gr.Column():
         gr.Markdown(constants.TOPICS_MD)

requirements.txt

@@ -1,4 +1,4 @@
-pip>=23.3
+pip>=24.1.1
 requests==2.31.0
 pylance==0.5.10
 pandas==2.0.3
@@ -9,8 +9,8 @@ scikit-learn==1.2.2
 plotly.express==0.4.1 
 openai==0.27.8
 openapi-schema-pydantic==1.2.4
-gradio==3.48.0
-gradio_client==0.6.1
+gradio==4.37.2
+gradio_client==1.0.2
 GitPython==3.1.31
 elevenlabs==0.2.18
 python-dotenv==1.0.0
@@ -25,4 +25,13 @@ pdfplumber==0.10.2
 pytz==2023.3
 easyocr==1.7.1
 ffmpeg==1.4
-spacy==3.7.2
+spacy==3.7.5
+PyMuPDF==1.24.7 
+PyMuPDFb==1.24.6
+pytesseract==0.3.10
+pdf2image==1.17.0
+Image==1.5.33 
+asgiref==3.8.1 
+django==5.0.6 
+sqlparse==0.5.0
+weasel==0.4.1

slack_sentiment_analysis/ui_sentiment_analysis.py

@@ -33,7 +33,7 @@ with gr.Blocks() as ui:
         with gr.Column():
             gr.Markdown(constants.DIRECTIONS_MD)
             gr.Markdown(value=constants.SLACK_SENTIMENT_ANALYSIS_DIRECTIONS)
-    code=gr.Textbox(label=constants.INITIAL_SLACK, interactive=True, placeholder=constants.SUMMARY_MD_DIRECTIONS, height="500", lines=3)   
+    code=gr.Textbox(label=constants.INITIAL_SLACK, interactive=True, placeholder=constants.SUMMARY_MD_DIRECTIONS, lines=3)   
     gr.Markdown(constants.SLACK_EXAMPLES_MD)
     with gr.Row():
         with gr.Column():

text_analysis/ui_text_analysis.py

@@ -48,8 +48,7 @@ with gr.Blocks() as ui:
         with gr.Row():
             with gr.Column():
                 inputString=gr.Textbox(placeholder=constants.NLP_PROMPT, 
-                                    label="Input Text", lines=3, height="500", 
-                                    interactive=True)
+                                    label="Input Text", lines=3,interactive=True)
                 with gr.Row():
                     clearBtn=gr.Button(constants.CLEAR, variant="secondary")
                     submitBtn=gr.Button(constants.SUBMIT, variant="primary")

utilities/constants.py

@@ -21,6 +21,7 @@ SLACK_SENTIMENT_ANALYSIS_DIRECTIONS="To get started select one of the slack data
 NLP_ANALYSIS_PURPOSE="Natural Language Processing (NLP) enables us to understand the structure and meaning in text. This view provides some understanding of NLP under the hood of Generative AI."
 NLP_ANALYSIS_DIRECTIONS="Start by entering a sentence or paragraph in the 'Text' box. Then tap 'Analyze'. You'll see token counts, part-of-speech tags, and a visual representation of the analysis."
 NLP_PROMPT="Enter text to analyze..."
+
 AR_ADVISOR_PURPOSE = "This component demonstrates how a chatbot can leverage local embeddings to provide context. In this example, embeddings were created from USAA's 2022 Member Report provide organizational context for all interactions. This shows how generative AI enables any dataset to be queried and explored via natural language."
 AR_ADVISOR_DIRECTIONS = "To get started, simply start asking questions. This bot has been trainged on USAA's 2022 Member Report and can answer questions about any content from the report. You can ask questions like: 'What's the best thing that happened last year?' or 'What was the biggest obstancle faced in 2022?'"
 
@@ -117,9 +118,8 @@ UI_5="Slack Sentiment Analysis"
 UI_6="Image to Text"
 UI_7="PDF Query Chatbot"
 UI_2="NLP Text Analysis"
+UI_9="AI Browser"
 
-HR_BENEFIT_PKL = 'utilities/data/embeddings/hr_asst_embeddings.pkl'
-HR_BENEFIT_CSV = 'utilities/data/embeddings/hr_asst_data.csv'
 
 PURPOSE_MD="### Purpose"
 DIRECTIONS_MD="### Directions"
@@ -220,7 +220,21 @@ CHUNK_LENGTH=2500
 #simple Chatbot
 PDF_INPUT_PATH = 'utilities/data/pdfs/usaa.pdf'
 TXT_OUTPUT_PATH = 'utilities/data/txt/usaa.txt'
+GIT_TXT_OUTPUT_PATH = 'utilities/data/txt/git.txt'
 
 CSV_PATH = 'utilities/data/csv/usaa.csv'
 PKL_PATH = 'utilities/data/pkl/usaa.pkl'
-EMBEDDING_CHUNK_LENGTH=500
+ZIP_INPUT_PATH='utilities/data/zip/main.zip'
+GIT_CSV_PATH = 'utilities/data/csv/git.csv'
+GIT_PKL_PATH = 'utilities/data/pkl/git.pkl'
+EMBEDDING_CHUNK_LENGTH=500
+
+
+HR_BENEFIT_PKL = 'utilities/data/embeddings/hr_asst_embeddings.pkl'
+HR_BENEFIT_CSV = 'utilities/data/embeddings/hr_asst_data.csv'
+
+POLICY_INPUT='utilities/data/pdfs/policy.pdf'
+POLICY_OUTPUT='utilities/data/csv/policy.csv'
+
+POLICY_TXT_PATH = 'utilities/data/txt/policy.txt'
+POLICY_PKL_PATH='utilities/data/pkl/policy.pkl'

utilities/data/embeddings/hr_asst_embeddings.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2fc7f21f90a8b7805ce45ed06780f97133cbf95faaf92d7655ee87de1ef8d16b
+oid sha256:6fbf6b6d88ea054dbd573a77207596b97f6e1274088d1ee005f9f55def73e714
 size 514467

utilities/data/pkl/usaa.pkl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:20bb2d36fe68b373ae234dc80ecd96a2f1bdd458034b80403cebef5865c2c65a
-size 1677861
+oid sha256:05815a4f9f1f7156b85cc1880995c9ceb5eb71730c78d703962b7bef388579d9
+size 1362488

utilities/prompt_constants.py

@@ -6,6 +6,7 @@ IT_EXPERT_PROMPT= "You are an IT support person. You are impatient and bored. Do
 INSURANCE_AGENT_PROMPT= "You are an insurance agent. Don't say your an AI language model and don't specify role in response. Your purpose is to ask questions in a turn-by-turn style to understand the situation. Ask for documents and policy numbers, simulate consulting insurance plans,keep customers informed with coverage and recovery option details. Along the way provide things like damage estimates, repair vendors, etc. Your responses are short and precise to gather information and provide information. Keep your responses short."
 INVESTIGATOR_PROMPT= "You are an investigator like Sherlock Holmes. Don't say your an AI language model and don't specify role in response. You job is solve a mystery. You do this by encouraging users to provide details you need. As you investigate, ask only one question at a time. You want to find things like witnesses, times of day, key locations, and event details. You should be professional and stoic, ask or answer one question at a time. Keep answers short."
 USAA_AR_EXPERT_PROMPT= "You represent USAA. Don't say your an AI language model and don't specify role. Provide answers to questions with kindness and respect. Be professional and warm, keep answers short."
+GIT_EXPERT_PROMPT= "You are a code anayzer. Don't say your an AI language model and don't specify role. Assume questions are computer science/software engineering in nature."
 
 #language presets
 SPANISH_PROMPT="Provide responses entirely in Spanish."