Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-10-24 23:38:55 -04:00
Code Issues Packages Projects Releases Wiki Activity

Update NYTimes

Browse Source 
Now they apparently are incapable of generating valid JSON even. Sigh.
This commit is contained in:
Kovid Goyal 2025-09-07 18:46:39 +05:30
parent cff6497bec
commit 921d82e63f
No known key found for this signature in database
GPG Key ID: 06BC317B515ACE7C
4 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
recipes/nytimes.recipe
View File

@ -267,7 +267,7 @@ def preloaded_data(soup):
raw = clean_js_json(raw) raw = clean_js_json(raw)
# with open('/t/raw.json', 'w') as f: # with open('/t/raw.json', 'w') as f:
# f.write(raw) # f.write(raw)
return json.loads(raw)['initialState'] return json.JSONDecoder(strict=False).raw_decode(raw)[0]['initialState']
def asset_to_article(asset): def asset_to_article(asset):

2
recipes/nytimes_sub.recipe
View File

@ -267,7 +267,7 @@ def preloaded_data(soup):
raw = clean_js_json(raw) raw = clean_js_json(raw)
# with open('/t/raw.json', 'w') as f: # with open('/t/raw.json', 'w') as f:
# f.write(raw) # f.write(raw)
return json.loads(raw)['initialState'] return json.JSONDecoder(strict=False).raw_decode(raw)[0]['initialState']
def asset_to_article(asset): def asset_to_article(asset):

2
recipes/nytimesbook.recipe
View File

@ -132,7 +132,7 @@ def preloaded_data(soup):
script = str(script) script = str(script)
raw = script[script.find('{') : script.rfind(';')].strip().rstrip(';') # } raw = script[script.find('{') : script.rfind(';')].strip().rstrip(';') # }
raw = clean_js_json(raw) raw = clean_js_json(raw)
return json.loads(raw)['initialState'] return json.JSONDecoder(strict=False).raw_decode(raw)[0]['initialState']
def parse_toc(soup): def parse_toc(soup):

4
src/calibre/web/site_parsers/nytimes.py
View File

@ -9,7 +9,7 @@ from xml.sax.saxutils import escape, quoteattr
from calibre.utils.iso8601 import parse_iso8601 from calibre.utils.iso8601 import parse_iso8601
module_version = 13 # needed for live updates module_version = 14 # needed for live updates
pprint pprint
@ -207,7 +207,7 @@ def clean_js_json(text):
def json_to_html(raw): def json_to_html(raw):
data = json.loads(clean_js_json(raw)) data = json.JSONDecoder(strict=False).raw_decode(raw)[0]
# open('/t/raw.json', 'w').write(json.dumps(data, indent=2)) # open('/t/raw.json', 'w').write(json.dumps(data, indent=2))
try: try:
data = data['initialData']['data'] data = data['initialData']['data']