Initial commit

app.py

@@ -0,0 +1,179 @@
+
+
+import warnings
+warnings.filterwarnings('ignore')
+
+from crewai import Agent, Task, Crew
+from crewai_tools import (
+    #DirectoryReadTool,
+    #FileReadTool,
+    #SerperDevTool,
+    #WebsiteSearchTool,
+    #DOCXSearchTool,
+    #RagTool,
+    TXTSearchTool
+)
+from datetime import datetime
+
+
+import os
+from utils import get_openai_api_key
+
+
+openai_api_key = get_openai_api_key()
+
+os.environ["OPENAI_MODEL_NAME"] = 'gpt-3.5-turbo'
+os.environ["OPENAI_API_KEY"] = "sk-proj-8nKBOgxInYCIqidcdSu7T3BlbkFJWac8qZbpFOE2TSn0OpId"
+
+
+def call_crew_kickoff(str_current_datetime):
+     # Instantiate tools
+     #meeting_trans_docs_tool = DirectoryReadTool(directory='./meeting-transcription')
+     #brd_temp_docs_tool = DirectoryReadTool(directory='./brd-template')
+     #file_tool = FileReadTool()
+     #web_rag_tool = WebsiteSearchTool()
+     #docx_search_tool = DOCXSearchTool()
+     #rag_tool = RagTool()
+     mt_tool = TXTSearchTool(txt='./meeting-transcription/meeting-transcript_' + str_current_datetime + '.txt')
+     brd_tool = TXTSearchTool(txt='./brd-template/brd-template.txt')
+
+     text_parsing_agent = Agent(
+        role="Text Interpreter",
+        goal="Parse and interpret the raw text input, structuring it into manageable sections"
+            "or data points that are relevant for analysis and processing.",
+        backstory="You excel at deciphering complex textual data. You act as the first line of analysis,"
+            "turning unstructured text into organized segments. You should enhance efficiency in data"
+            "handling and support subsequent stages of data processing.",
+        tools=[mt_tool],
+        allow_delegation=True,
+            verbose=True
+     )
+
+     data_extraction_agent = Agent(
+          role="Data Extraction Agent",
+          goal="Identify and extract essential data points, statistics,"
+               "and specific information from the parsed text that are crucial"
+               "for drafting a Business Requirements Document.",
+          backstory="You should tackle the challenge of sifting through detailed textual data to"
+               "find relevant information. You should be doing it with precision and speed, equipped"
+               "with capabilities to recognize and categorize data efficiently, making it invaluable"
+               "for projects requiring quick turnaround and accurate data handling.",
+          tools=[mt_tool, brd_tool],
+          allow_delegation=True,
+          verbose=True
+     )
+
+     brd_compiler_agent = Agent(
+          role="BRD Compiler",
+          goal="Assemble the extracted data into a well-structured Business Requirements Document,"
+               "ensuring that it is clear, coherent, and formatted according to standards.",
+          backstory="You are a meticulous Business Requirement Document compiler, You should alleviate"
+               "the burdens of manual document assembly. Ensure that all documents are crafted with"
+               "precision, adhering to organizational standards, and ready for stakeholder review. You"
+               "should be automating routine documentation tasks, thus allowing human team members to focus"
+               "on more strategic activities.",
+          tools=[brd_tool],
+          allow_delegation=True,
+          verbose=True
+     )
+
+     text_parsing = Task(
+        description=(
+            "1. Open and read the contents of the input text file.\n"
+            "2. Analyze the document structure to identify headings, subheadings, and key paragraphs.\n"
+            "3. Extract text under each identified section, ensuring context is preserved.\n"
+            "4. Format the extracted text into a JSON structure with labels indicating the type"
+                "of content (e.g., heading, detail)."
+        ),
+        expected_output="Structured JSON object containing separated sections of"
+                        "text with labels based on their content type.",
+        agent=text_parsing_agent,
+     )
+
+     data_extraction = Task(
+        description=(
+            "1. Take the JSON structured data from the Text Parsing Agent.\n"
+            "2. Identify and extract specific data points like project goals, technical requirements,"
+                "and stakeholder information.\n"
+            "3. Organize the extracted data into relevant categories for easy access and use.\n"
+            "4. Format all extracted data into a structured form suitable for document generation,"
+                "ensuring it's ready for template insertion.\n"
+        ),
+        expected_output="A comprehensive list of key data points organized by category, ready for use in document generation.",
+        agent=data_extraction_agent,
+     )
+
+     compile_brd = Task(
+        description=(
+            "1. Accept the structured and categorized data from the Data Extraction Agent.\n"
+            "2. Open and read the BRD template for data insertion.\n"
+            "3. Insert the received data into the respective sections of the BRD template.\n"
+            "4. Apply formatting rules to ensure the document is professional and adheres to standards.\n"
+            "5. Save the populated and formatted document as a new markdown file, marking the task as complete.\n"
+        ),
+        expected_output="A complete Business Requirements Document in markdown format, ready for review and distribution.",
+        agent=brd_compiler_agent,
+        output_file='generated-brd/brd_' + str_current_datetime + '.md',  # The final blog post will be saved here
+     )
+
+     crew = Crew(
+        agents=[text_parsing_agent, data_extraction_agent, brd_compiler_agent],
+        tasks=[text_parsing, data_extraction, compile_brd],
+        verbose=2
+     )
+
+     result = crew.kickoff(inputs={'datetime': str_current_datetime})
+
+
+import gradio as gr
+
+def process_file(input_file):
+    current_datetime = datetime.now().strftime("%Y-%m-%d %H-%M-%S")
+    print("Current date & time : ", current_datetime)
+    
+    # convert datetime obj to string
+    str_current_datetime = str(current_datetime)
+
+    fh = open(input_file, 'rb')
+    #data = fh.read()
+
+    # Ensure the target directory exists
+    output_dir = "meeting-transcription"
+    if not os.path.exists(output_dir):
+        os.makedirs(output_dir)
+    
+    # Save the uploaded file to the specified folder with the specified name
+    input_filepath = os.path.join(output_dir, "meeting-transcript_" + str_current_datetime + ".txt")
+    with open(input_filepath, "wb") as file:
+        file.write(fh.read())
+    fh.close()
+
+    call_crew_kickoff(str_current_datetime)
+
+    # Example of processing: adding Markdown formatting. Replace this with your actual processing.
+    #processed_text = "# Processed Output\n\n" + "\n".join(f"- {line}" for line in text.splitlines())
+    
+    output_filename = "generated-brd/brd_" + str_current_datetime + ".md"
+    #with open(output_filename, "w") as file:
+    #    file.write(processed_text)
+
+    # Read the contents of the generated Markdown file
+    with open(output_filename, "r") as md_file:
+        markdown_content = md_file.read()
+
+    # Return both the filename for download and the Markdown content for display
+    return output_filename, markdown_content
+
+with gr.Blocks() as demo:
+    with gr.Row():
+        file_input = gr.File(label="Upload a text file")
+        download_btn = gr.File(label="Download Processed File in Markdown", file_count="single")
+    with gr.Row():
+        markdown_output = gr.Markdown()
+
+    file_input.change(process_file, inputs=file_input, outputs=[download_btn, markdown_output])
+
+demo.launch(share=True)
+
+
+