WIP - Streamlit UI, Porting CLI

This commit is contained in:
ajaysi
2024-06-02 23:05:27 +05:30
parent ae8c9d0ac3
commit 2a1bb49020
11 changed files with 645 additions and 884 deletions

View File

@@ -1,7 +1,7 @@
import os
import sys
import json
import configparser
from pathlib import Path
from loguru import logger
logger.remove()
@@ -13,26 +13,27 @@ logger.add(sys.stdout,
from ..gpt_providers.text_generation.main_text_generation import llm_text_gen
# FIXME: Provide num_blogs, num_faqs as inputs.
def write_blog_google_serp(search_keyword, search_results):
"""Combine the given online research and gpt blog content"""
"""Combine the given online research and GPT blog content"""
try:
config_path = os.path.abspath(os.path.join(os.path.dirname(__file__), '..', '..', 'main_config'))
config = configparser.ConfigParser()
config.read(config_path, encoding='utf-8')
config_path = Path(os.environ["ALWRITY_CONFIG"])
with open(config_path, 'r', encoding='utf-8') as file:
config = json.load(file)
except Exception as err:
print(f"Error: Failed to read values from config: {err}")
logger.error(f"Error: Failed to read values from config: {err}")
exit(1)
blog_characteristics = config['Blog Content Characteristics']
prompt = f"""
As expert Creative Content writer,
I want you to write {config.get('blog_characteristics', 'blog_type')} blog post,
I want you to write {blog_characteristics['Blog Type']} blog post,
that explores {search_keyword} and also include 5 FAQs.
Below are the guidelines to follow:
1). You must respond in {config.get('blog_characteristics', 'blog_language')} language.
2). Tone and Brand Alignment: Adjust your tone, voice, personality for {config.get('blog_characteristics', 'blog_tone')} audience.
3). Make sure your response content length is of {config.get('blog_characteristics', 'blog_length')} words.
1). You must respond in {blog_characteristics['Blog Language']} language.
2). Tone and Brand Alignment: Adjust your tone, voice, personality for {blog_characteristics['Blog Tone']} audience.
3). Make sure your response content length is of {blog_characteristics['Blog Length']} words.
4). Include FAQs from 'People also Ask' section of provided context 'google search result'.
I want the post to offer unique insights, relatable examples, and a fresh perspective on the topic.
@@ -40,7 +41,9 @@ def write_blog_google_serp(search_keyword, search_results):
\n\n
\"\"\"{search_results}\"\"\"
"""
logger.info("Generating blog and FAQs from Google web search results.")
try:
response = llm_text_gen(prompt)
return response
@@ -48,7 +51,6 @@ def write_blog_google_serp(search_keyword, search_results):
logger.error(f"Exit: Failed to get response from LLM: {err}")
exit(1)
def improve_blog_intro(blog_content, blog_intro):
"""Combine the given online research and gpt blog content"""
prompt = f"""

View File

@@ -44,9 +44,11 @@ def write_blog_from_keywords(search_keywords, url=None):
status.update(label=f"🛀 Starting Tavily AI research: {search_keywords}")
tavily_search_result, t_titles, t_answer = do_tavily_ai_search(search_keywords)
status.update(label=f"🙆 Finished Google Search & Tavily AI Search on: {search_keywords}", expanded=False)
status.update(label=f"🙆 Finished Google Search & Tavily AI Search on: {search_keywords}",
state="complete", expanded=False)
except Exception as err:
st.error(f"Failed in web research: {err}")
logger.error(f"Failed in web research: {err}")
with st.status("Started Writing blog from google search..", expanded=True) as status:
@@ -56,12 +58,9 @@ def write_blog_from_keywords(search_keywords, url=None):
status.update(label=f"🛀 Writing blog from Google Search on: {search_keywords}")
blog_markdown_str = write_blog_google_serp(search_keywords, google_search_result)
st.markdown(blog_markdown_str)
# Hate the robotic introductions.
#blog_markdown_str = improve_blog_intro(blog_markdown_str, t_answer)
#st.markdown(blog_markdown_str)
status.update(label="🙎 Draft 1: Your Content from Google search result.", expanded=False)
status.update(label="🙎 Draft 1: Your Content from Google search result.", state="complete", expanded=False)
except Exception as err:
st.error(f"Failed in Google web research: {err}")
logger.error(f"Failed in Google web research: {err}")
# logger.info/check the final blog content.

View File

@@ -60,22 +60,26 @@ def long_form_generator(content_keywords):
"""
with st.status("Start Writing Long Form Article, Hold my Beer..", expanded=True) as status:
# Read the main_config to define tone, character, personality of the content to be generated.
try:
try:
status.update(label=f"Starting to write content on {content_keywords}.")
logger.info(f"Starting to write content on {content_keywords}.")
# Define persona and writing guidelines
content_tone, target_audience, content_type, content_language, output_format = read_return_config_section('blog_characteristics')
content_tone, target_audience, content_type, content_language, output_format, content_length = read_return_config_section('blog_characteristics')
except Exception as err:
logger.error(f"Failed to Read config params from main_config: {err}")
return
st.error(f"Failed to Read config params from main_config: {err}")
return False
try:
filepath = os.path.join(os.environ["PROMPTS_DIR"], "long_form_ai_writer.prompts")
status.update(label=f"Reading Prompts from {filepath}.")
# Check if file exists
if not os.path.exists(filepath):
raise FileNotFoundError(f"File {filepath} does not exist")
with open(filepath, 'r') as file:
prompts = yaml.safe_load(file)
except Exception as err:
st.error(f"Exit: Failed to read prompts from {filepath}: {err}")
logger.error(f"Exit: Failed to read prompts from {filepath}: {err}")
exit(1)
@@ -147,11 +151,12 @@ def long_form_generator(content_keywords):
content_title=content_title,
web_research_result=web_research_result)).text
logger.info(f"The content Outline is: {content_outline}\n\n")
status.update(label="Generated the content outline.")
status.update(label=f"Completed with Content Outline.")
except Exception as err:
logger.error(f"Failed to generate content outline: {err}")
try:
status.update(label=f"Do web research with Tavily to provide context for content creation.")
logger.info("Do web research with Tavily to provide context for content creation.")
# Do Metaphor/Exa AI search.
table_data = []
@@ -163,6 +168,7 @@ def long_form_generator(content_keywords):
web_research_result = table_data
except Exception as err:
logger.error(f"Failed to do Tavily AI search: {err}")
st.error(f"Failed to do Tavily AI search: {err}")
return
try:
@@ -172,6 +178,7 @@ def long_form_generator(content_keywords):
web_research_result=web_research_result,
writing_guidelines=writing_guidelines)).text
except Exception as err:
st.error(f"Failed to Generate Starting draft: {err}")
logger.error(f"Failed to Generate Starting draft: {err}")
return
@@ -194,10 +201,11 @@ def long_form_generator(content_keywords):
logger.error(f"Failed as: {err} and {continuation}")
logger.info(f"Writing in progress... Current draft length: {len(draft)} characters")
status.update(label=f"Writing in progress... Current draft length: {len(draft)} characters")
search_terms = f"""
I will provide you with blog outline, your task is to read the outline & return 3 google search keywords.
Your response will be used to do web research for writing on the given outline.
Do not explain your response, provide 3 google search sentences encompassing the given content outline.
Do not explain your response, provide 8 google search sentences encompassing the given content outline.
Provide the search term results as comma separated values.\n\n
Content Outline:\n
'{content_outline}'
@@ -227,26 +235,36 @@ def long_form_generator(content_keywords):
# At this point, the context is little stale. We should more web research on
# related queries as per the content outline, to augment the LLM context.
except Exception as err:
st.error(f"Failed to continually write the Essay: {err}")
logger.error(f"Failed to continually write the Essay: {err}")
return
# Remove 'IAMDONE' and print the final story
final = draft.replace('IAMDONE', '').strip()
blog_title, blog_meta_desc, blog_tags, blog_categories = blog_metadata(final,
content_keywords, m_titles)
generated_image_filepath = None
# TBD: Save the blog content as a .md file. Markdown or HTML ?
save_blog_to_file(final, blog_title, blog_meta_desc, blog_tags, blog_categories, generated_image_filepath)
blog_frontmatter = f"""
---
title: {blog_title}
categories: [{blog_categories}]
tags: [{blog_tags}]
Meta description: {blog_meta_desc.replace(":", "-")}
---"""
logger.info(f"\n{blog_frontmatter}{final}\n\n")
st.write(f"\n{blog_frontmatter}{final}\n\n")
status.update(label="Success: Finished writing Long form content.")
# FIXME: The current implementation is suited for normal length content.
# In long content sending the whole content for each content metadata is expensive.
# blog_title, blog_meta_desc, blog_tags, blog_categories = blog_metadata(final,
# content_keywords, m_titles)
# status.update(label="Success: Finished with Title, Meta Description, Tags, categories")
# generated_image_filepath = None
# # TBD: Save the blog content as a .md file. Markdown or HTML ?
# save_blog_to_file(final, blog_title, blog_meta_desc, blog_tags, blog_categories, generated_image_filepath)
#
# blog_frontmatter = dedent(f"""
# \n---------------------------------------------------------------------
# title: {blog_title.strip()}\n
# categories: [{blog_categories.strip()}]\n
# tags: [{blog_tags.strip()}]\n
# Meta description: {blog_meta_desc.replace(":", "-").strip()}\n
# ---------------------------------------------------------------------\n
# """)
#
# logger.info(f"\n{blog_frontmatter}{final}\n\n")
# st.markdown(f"\n{blog_frontmatter}{final}\n\n")
logger.info(f"\n{final}\n\n")
logger.info(f"\n\n ################ Finished writing Blog for : {content_keywords} #################### \n")
with st.expander("**Click to View the final content draft:**"):
st.markdown(f"\n{final}\n\n")