kunthawat/moreminimore-marketing
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Base code

Browse Source
This commit is contained in:
Kunthawat Greethong
2026-01-08 22:39:53 +07:00
parent 697115c61a
commit c35fa52117
2169 changed files with 626670 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

202
ToBeMigrated/ai_writers/scholar_blogs/main_arxiv_to_blog.py Normal file
View File

@@ -0,0 +1,202 @@
import sys
import os
import datetime
import tiktoken
from .arxiv_schlorly_research import fetch_arxiv_data, create_dataframe, get_arxiv_main_content
from .arxiv_schlorly_research import arxiv_bibtex, scrape_images_from_arxiv, download_image
from .arxiv_schlorly_research import read_written_ids, extract_arxiv_ids_from_line, append_id_to_file
from .write_research_review_blog import review_research_paper
from .combine_research_and_blog import blog_with_research
from .write_blog_scholar_paper import write_blog_from_paper
from .gpt_providers.gemini_pro_text import gemini_text_response
from .generate_image_from_prompt import generate_image
from .convert_content_to_markdown import convert_tomarkdown_format
from .get_blog_metadata import blog_metadata
from .get_code_examples import gemini_get_code_samples
from .save_blog_to_file import save_blog_to_file
from .take_url_screenshot import screenshot_api
from loguru import logger
logger.remove()
logger.add(sys.stdout,
colorize=True,
format="<level>{level}</level>|<green>{file}:{line}:{function}</green>| {message}"
)
def blog_arxiv_keyword(query):
""" Write blog on given arxiv paper."""
arxiv_id = None
arxiv_url = None
bibtex = None
research_review = None
column_names = ['Title', 'Date', 'Id', 'Summary', 'PDF URL']
papers = fetch_arxiv_data(query)
df = create_dataframe(papers, column_names)
for paper in papers:
# Extracting the arxiv_id
arxiv_id = paper[2].split('/')[-1]
arxiv_url = "https://browse.arxiv.org/html/" + arxiv_id
bibtex = arxiv_bibtex(arxiv_id)
logger.info(f"Get research paper text from the url: {arxiv_url}")
research_content = get_arxiv_main_content(arxiv_url)
num_tokens = num_tokens_from_string(research_content, "cl100k_base")
logger.info(f"Number of tokens sent: {num_tokens}")
# If the number of tokens is below the threshold, process and print the review
if 1000 < num_tokens < 30000:
logger.info(f"Writing research review on {paper[0]}")
research_review = review_research_paper(research_content)
research_review = f"\n{research_review}\n\n" + f"```{bibtex}```"
#research_review = research_review + "\n\n\n" + f"{df.to_markdown()}"
research_review = convert_tomarkdown_format(research_review, "gemini")
break
else:
# Skip to the next iteration if the condition is not met
continue
logger.info(f"Final scholar article: \n\n{research_review}\n")
# TBD: Scrape images from research reports and pass to vision to get conclusions out of it.
#image_urls = scrape_images_from_arxiv(arxiv_url)
#print("Downloading images found on the page:")
#for img_url in image_urls:
# download_image(img_url, arxiv_url)
try:
blog_postprocessing(arxiv_id, research_review)
except Exception as err:
logger.error(f"Failed in blog post processing: {err}")
sys.exit(1)
logger.info(f"\n\n ################ Finished writing Blog for : #################### \n")
def blog_arxiv_url_list(file_path):
""" Write blogs on all the arxiv links given in a file. """
extracted_ids = []
try:
with open(file_path, 'r', encoding="utf-8") as file:
for line in file:
arxiv_id = extract_arxiv_ids_from_line(line)
if arxiv_id:
extracted_ids.append(arxiv_id)
except FileNotFoundError:
logger.error(f"File not found: {file_path}")
raise FileNotFoundError
except Exception as e:
logger.error(f"Error while reading the file: {e}")
raise e
# Read already written IDs
written_ids = read_written_ids('papers_already_written_on.txt')
# Loop through extracted IDs
for arxiv_id in extracted_ids:
if arxiv_id not in written_ids:
# This ID has not been written on yet
arxiv_url = "https://browse.arxiv.org/html/" + arxiv_id
logger.info(f"Get research paper text from the url: {arxiv_url}")
research_content = get_arxiv_main_content(arxiv_url)
try:
num_tokens = num_tokens_from_string(research_content, "cl100k_base")
except Exception as err:
logger.error(f"Failed in counting tokens: {err}")
sys.exit(1)
logger.info(f"Number of tokens sent: {num_tokens}")
# If the number of tokens is below the threshold, process and print the review
# FIXME: Docs over 30k tokens, need to be chunked and summarized.
if 1000 < num_tokens < 30000:
try:
logger.info(f"Getting bibtex for arxiv ID: {arxiv_id}")
bibtex = arxiv_bibtex(arxiv_id)
except Exception as err:
logger.error(f"Failed to get Bibtex: {err}")
try:
logger.info(f"Writing a research review..")
research_review = review_research_paper(research_content, "gemini")
logger.info(f"Research Review: \n{research_review}\n\n")
except Exception as err:
logger.error(f"Failed to write review on research paper: {arxiv_id}{err}")
research_blog = write_blog_from_paper(research_content, "gemini")
logger.info(f"\n\nResearch Blog: {research_blog}\n\n")
research_blog = f"\n{research_review}\n\n" + f"```\n{bibtex}\n```"
#research_review = blog_with_research(research_review, research_blog, "gemini")
#logger.info(f"\n\n\nBLOG_WITH_RESEARCh: {research_review}\n\n\n")
research_review = convert_tomarkdown_format(research_review, "gemini")
research_review = f"\n{research_review}\n\n" + f"```{bibtex}```"
logger.info(f"Final blog from research paper: \n\n{research_review}\n\n\n")
try:
blog_postprocessing(arxiv_id, research_review)
except Exception as err:
logger.error(f"Failed in blog post processing: {err}")
sys.exit(1)
logger.info(f"\n\n ################ Finished writing Blog for : #################### \n")
else:
# Skip to the next iteration if the condition is not met
logger.error("FIXME: Docs over 30k tokens, need to be chunked and summarized.")
continue
else:
logger.warning(f"Already written, skip writing on Arxiv paper ID: {arxiv_id}")
def blog_postprocessing(arxiv_id, research_review):
""" Common function to do blog postprocessing. """
try:
append_id_to_file(arxiv_id, "papers_already_written_on.txt")
except Exception as err:
logger.error(f"Failed to write/append ID to papers_already_written_on.txt: {err}")
raise err
try:
blog_title, blog_meta_desc, blog_tags, blog_categories = blog_metadata(research_review)
except Exception as err:
logger.error(f"Failed to get blog metadata: {err}")
raise err
try:
arxiv_url_scrnsht = f"https://arxiv.org/abs/{arxiv_id}"
generated_image_filepath = take_paper_screenshot(arxiv_url_scrnsht)
except Exception as err:
logger.error(f"Failed to tsk paper screenshot: {err}")
raise err
try:
save_blog_to_file(research_review, blog_title, blog_meta_desc, blog_tags,\
blog_categories, generated_image_filepath)
except Exception as err:
logger.error(f"Failed to save blog to a file: {err}")
sys.exit(1)
def take_paper_screenshot(arxiv_url):
""" Common function to take paper screenshot. """
# fixme: Remove the hardcoding, need add another option OR in config ?
image_dir = os.path.join(os.getcwd(), "blog_images")
generated_image_name = f"generated_image_{datetime.datetime.now():%Y-%m-%d-%H-%M-%S}.png"
generated_image_filepath = os.path.join(image_dir, generated_image_name)
if arxiv_url:
try:
generated_image_filepath = screenshot_api(arxiv_url, generated_image_filepath)
except Exception as err:
logger.error(f"Failed in taking url screenshot: {err}")
return generated_image_filepath
def num_tokens_from_string(string, encoding_name):
"""Returns the number of tokens in a text string."""
try:
encoding = tiktoken.get_encoding(encoding_name)
num_tokens = len(encoding.encode(string))
return num_tokens
except Exception as err:
logger.error(f"Failed to count tokens: {err}")
sys.exit(1)

49
ToBeMigrated/ai_writers/scholar_blogs/write_blog_scholar_paper.py Normal file
View File

@@ -0,0 +1,49 @@
import sys
from .gpt_providers.openai_chat_completion import openai_chatgpt
from .gpt_providers.gemini_pro_text import gemini_text_response
from loguru import logger
logger.remove()
logger.add(sys.stdout,
colorize=True,
format="<level>{level}</level>|<green>{file}:{line}:{function}</green>| {message}"
)
def write_blog_from_paper(paper_content):
""" Write blog from given paper url. """
prompt = f"""As an expert in NLP and AI, I will provide you with a content of a research paper.
Your task is to write a highly detailed blog(at least 2000 words), breaking down complex concepts for beginners.
Take your time and do not rush to respond.
Do not provide explanations, suggestions in your response.
Include the below section in your blog:
Highlights: Include a list of 5 most important and unique claims of the given research paper.
Abstract: Start by reading the abstract, which provides a concise summary of the research, including its purpose, methodology, and key findings.
Introduction: This section will give you background information and set the context for the research. It often ends with a statement of the research question or hypothesis.
Methodology: Include description of how authors conducted the research. This can include data sources, experimental setup, analytical techniques, etc.
Results: This section presents the data or findings of the research. Pay attention to figures, tables, and any statistical analysis provided.
Discussion/Analysis: In this section, Explain how research paper answers the research questions or how they fit with existing knowledge.
Conclusion: This part summarizes the main findings and their implications. It might also suggest areas for further research.
References: The cited works can provide additional context or background reading.
Remember, Please use MLA format and markdown syntax.
Do not provide description, explanations for your response.
Take your time in crafting your blog content, do not rush to give the response.
Using the blog structure above, please write a detailed and original blog on given research paper: \n'{paper_content}'\n\n"""
if 'gemini' in gpt_providers:
try:
response = gemini_text_response(prompt)
return response
except Exception as err:
logger.error(f"Failed to get response from gemini: {err}")
raise err
elif 'openai' in gpt_providers:
try:
logger.info("Calling OpenAI LLM.")
response = openai_chatgpt(prompt)
return response
except Exception as err:
logger.error(f"failed to get response from Openai: {err}")
raise err

89
ToBeMigrated/ai_writers/scholar_blogs/write_research_review_blog.py Normal file
View File

@@ -0,0 +1,89 @@
import sys
from .gpt_providers.openai_chat_completion import openai_chatgpt
from .gpt_providers.gemini_pro_text import gemini_text_response
from .gpt_providers.mistral_chat_completion import mistral_text_response
from loguru import logger
logger.remove()
logger.add(sys.stdout,
colorize=True,
format="<level>{level}</level>|<green>{file}:{line}:{function}</green>| {message}"
)
def review_research_paper(research_blog):
""" """
prompt = f"""As world's top researcher and academician, I will provide you with research paper.
Your task is to write a highly detailed review report.
Important, your report should be factual, original and demostrate your expertise.
Review guidelines:
1). Read the Abstract and Introduction Carefully:
Begin by thoroughly reading the abstract and introduction of the paper.
Try to understand the research question, the objectives, and the background information.
Identify the central argument or hypothesis that the study is examining.
2). Examine the Methodology and Methods:
Read closely at the research design, whether it is experimental, observational, qualitative, or a combination of methods.
Check the sampling strategy and the size of the sample.
Review the methods of data collection and the instruments used for this purpose.
Think about any ethical issues and possible biases in the study.
3). Analyze the Results and Discussion:
Review how the results are presented, including any tables, graphs, and statistical analysis.
Evaluate the findings' validity and reliability.
Analyze whether the results support or contradict the research question and hypothesis.
Read the discussion section where the authors interpret their findings and their significance.
4). Consider the Limitations and Strengths:
Spot any limitations or potential weaknesses in the study.
Evaluate the strengths and contributions that the research makes.
Think about how generalizable the findings are to other populations or situations.
5). Assess the Writing and Organization:
Judge the clarity and structure of the report.
Consider the use of language, grammar, and the overall formatting.
Assess how well the arguments are logically organized and how coherent the report is.
6). Evaluate the Literature Review:
Examine how comprehensive and relevant the literature review is.
Consider how the study adds to or builds upon existing research.
Evaluate the timeliness and quality of the sources cited in the research.
7). Review the Conclusion and Implications:
Look at the conclusions drawn from the study and how well they align with the findings.
Think about the practical implications and potential applications of the research.
Evaluate the suggestions for further research or policy actions.
8). Overall Assessment:
Formulate an overall opinion about the research report's quality and thoroughness.
Consider the significance and impact of the findings.
Evaluate how the study contributes to its field of research.
9). Provide Constructive Feedback:
Offer constructive criticism and suggestions for improvement, where necessary.
Think about possible biases or alternative ways to interpret the findings.
Suggest ideas for future research or for replicating the study.
Do not provide description, explanations for your response.
Using the above review guidelines, write a detailed review report on the below research paper.
Research Paper: '{research_blog}'
"""
if 'gemini' in gpt_providers:
try:
response = gemini_text_response(prompt)
return response
except Exception as err:
logger.error(f"Failed to get response from gemini: {err}")
response = mistral_text_response(prompt)
return response
elif 'openai' in gpt_providers:
try:
logger.info("Calling OpenAI LLM.")
response = openai_chatgpt(prompt)
return response
except Exception as err:
SystemError(f"Failed to get response from Openai: {err}")