gemini_request.py

import yaml
from google import genai
from google.genai import types
import datetime
import requests  # This may need to be installed from pip

import urllib3
# https://urllib3.readthedocs.io/en/latest/advanced-usage.html#tls-warnings
urllib3.disable_warnings()

def read_value_from_yaml(file_path, key):
  """Reads a value from a YAML file.

  Args:
    file_path: The path to the YAML file.
    key: The key to retrieve the value for.

  Returns:
    The value associated with the key, or None if the key is not found
    or if there's an error parsing the YAML file.
  """
  with open(file_path, 'r') as f:
    data = yaml.safe_load(f)  # Use safe_load for security
  if isinstance(data, dict) and key in data:
    return data[key]
  return None

def fetch_reader_document_list_api(updated_after=None, location=None, category=None, content=True, token=None):
    full_data = []
    next_page_cursor = None
    while True:
        params = {}
        if next_page_cursor:
            params['pageCursor'] = next_page_cursor
        if updated_after:
            params['updatedAfter'] = updated_after
        if location:
            params['location'] = location
        if category:
            params['category'] = category
        if content:
            params['withHtmlContent'] = content

        ## print("Making export api request with params " + str(params) + "...")
        response = requests.get(
            url="https://readwise.io/api/v3/list/",
            params=params,
            headers={"Authorization": f"Token {token}"}, verify=False
        )
        try:
            full_data.extend(response.json()['results'])
        except KeyError:
            print(response.json()); exit()

        next_page_cursor = response.json().get('nextPageCursor')
        if not next_page_cursor:
            break
    return full_data

gkey = read_value_from_yaml('secret.yaml', 'galf')
rkey = read_value_from_yaml('secret.yaml', 'token')

# Later, if you want to get new documents updated after some date, do this:
docs_after_date = datetime.datetime.now() - datetime.timedelta(days=7)  # use your own stored date
print(docs_after_date)
article_delta = fetch_reader_document_list_api(updated_after=docs_after_date.isoformat(), token=rkey)

client = genai.Client(api_key=gkey)

articles = [str(x) for x in article_delta]

print("Article count is {}".format(len(articles)))

system_instruction1 = """
Role: You are an assistant summarizing articles that are of interest.
Focus on newly introduced techniques, tools, or approaches especially as
they relate to information security
Make the title a link to the source material starting the title with the top level domain linked and including the article name separated by a colon
Include an entry for every article
Include meta commentary on common sources and topics
Render output as stand alone valid HTML without any other intro
"""

system_instruction = """
Role: You are an executive assistant

Using this structured set of article metadata extract common topics and key insights
across all articles with reference links to the source web page.  Make the article title the link to the article with
a one sentence summary to the right.

Break into non-technology, cybersecurity, non-cybersecurity technology, Amazon Web Services (AWS) sections with a few sentences of summary of key highlights
with each article only appearing in one section related with a special section at the bottom enumerating
tools and products with a link and brief description. 

Format this in HTML and without any other intro.

Let me know if a site is an especially popular source of references and
anything else that seems interesting and unique across articles

Be only as verbose as you have room based on max_output_tokens to not leave out any article entirely.
Truncate line length at 12 characters in the HTML.

"""


response = client.models.generate_content(
    model="gemini-2.0-flash",
    config=types.GenerateContentConfig(
        system_instruction=system_instruction,
        max_output_tokens=4000,
        temperature=0.1),
    contents=articles,
)

# cleanup html
text = response.text.rstrip("```").lstrip("```").lstrip('html')

filepath = 'page.html'
try:
  with open(filepath, 'w', encoding='utf-8') as f:
      f.write(text)
      print(f"Successfully wrote to {filepath}")
except Exception as e:
    print(f"Error writing to {filepath}: {e}")

# print(response)