sij/sijapi
1
1
Fork 0
Code Issues3 Pull requests Projects Releases Packages Wiki Activity Actions

Auto-update: Fri Jan 17 11:58:21 PST 2025

Browse source
This commit is contained in:
sanj 2025-01-17 11:58:21 -08:00
parent 1863edbc06
commit ab86c3566f
1 changed files with 5 additions and 9 deletions

14
sijapi/routers/news.py
View file

@ -20,13 +20,12 @@ from fastapi import APIRouter, BackgroundTasks, UploadFile, Form, HTTPException,
from pathlib import Path from pathlib import Path
from sijapi import Archivist, News, Tts, OBSIDIAN_VAULT_DIR, OBSIDIAN_RESOURCES_DIR from sijapi import Archivist, News, Tts, OBSIDIAN_VAULT_DIR, OBSIDIAN_RESOURCES_DIR
from sijapi.utilities import html_to_markdown, download_file, sanitize_filename, assemble_journal_path, assemble_archive_path, contains_profanity, is_ad_or_tracker from sijapi.utilities import html_to_markdown, download_file, sanitize_filename, assemble_journal_path, assemble_archive_path, contains_profanity, is_ad_or_tracker
from sijapi.routers import gis, llm, tts, note from sijapi.routers import gis, llm, tts, note, email
from sijapi.logs import get_logger from sijapi.logs import get_logger
l = get_logger(__name__) l = get_logger(__name__)
news = APIRouter() news = APIRouter()
@news.post("/clip") @news.post("/clip")
async def clip_post( async def clip_post(
bg_tasks: BackgroundTasks, bg_tasks: BackgroundTasks,
@ -153,7 +152,8 @@ async def process_and_save_article(
title: Optional[str] = None, title: Optional[str] = None,
tts_mode: str = "summary", tts_mode: str = "summary",
voice: str = Tts.elevenlabs.default, voice: str = Tts.elevenlabs.default,
site_name: Optional[str] = None site_name: Optional[str] = None,
email_address: str = None
) -> str: ) -> str:
''' '''
Primary function for saving articles. Primary function for saving articles.
@ -162,7 +162,7 @@ Primary function for saving articles.
# Fetch and parse article # Fetch and parse article
article = await fetch_and_parse_article(url) article = await fetch_and_parse_article(url)
try: try:
# Generate title and file paths # Generate title and file paths
title = sanitize_filename(title or article.title or f"Untitled - {dt_datetime.now().strftime('%Y-%m-%d')}") title = sanitize_filename(title or article.title or f"Untitled - {dt_datetime.now().strftime('%Y-%m-%d')}")
markdown_filename, relative_path = assemble_journal_path(dt_datetime.now(), subdir="Articles", filename=title, extension=".md") markdown_filename, relative_path = assemble_journal_path(dt_datetime.now(), subdir="Articles", filename=title, extension=".md")
@ -197,10 +197,6 @@ Primary function for saving articles.
raise HTTPException(status_code=500, detail=str(e)) raise HTTPException(status_code=500, detail=str(e))
from newspaper import Article as NewspaperArticle
async def fetch_and_parse_article(url: str) -> Article: async def fetch_and_parse_article(url: str) -> Article:
# Try trafilatura first # Try trafilatura first
source = trafilatura.fetch_url(url) source = trafilatura.fetch_url(url)
@ -230,7 +226,7 @@ async def fetch_and_parse_article(url: str) -> Article:
# If trafilatura fails, use newspaper3k # If trafilatura fails, use newspaper3k
try: try:
newspaper_article = NewspaperArticle(url) newspaper_article = Article(url)
newspaper_article.download() newspaper_article.download()
newspaper_article.parse() newspaper_article.parse()