Repository navigation
Tags: ScrapeUnblocker/scrapeunblocker-python
Tags
Google Hotels, Google News, Ads Transparency, Shopee, Taobao and Upwo… …rk methods; 0.14.0 16 plugin methods generated from the public OpenAPI spec: google_hotels, google_hotels_property, google_news, google_ads_transparency, google_ads_transparency_advertisers, shopee_search, shopee_product, taobao_search, taobao_product and upwork_jobs, plus the Smyths Toys (smyths_inventory, smyths_details, smyths_search, smyths_category) and Wish (wish_search, wish_product) methods. The README plugin table lists them by category. Temu's plugin is unlisted, so the generator no longer emits temu_search and temu_product. Both move to the hand-written Client and AsyncClient with the same signatures and are marked deprecated, the way the TikTok and Southwest methods are kept; the README table no longer lists them.
get_markdown: the page as clean Markdown for LLMs; 0.13.0 getPageSource's new markdown=true returns the rendered page as clean Markdown (text/markdown) for LLMs, agents and RAG pipelines. The method sends main_content_only=false only when the caller turns it off (the API keeps only the main content by default) and takes the same render options as the HTML call. Billed like an HTML call; a target 404/410 maps to the existing TargetNotFound error with the Markdown page.
get_image: document original-format bytes and its 404/415 errors; 0.11.1 The API's /getImage now returns the image exactly as the host serves it instead of a PNG re-encode. A missing image is the host's 404/410 with X-Origin-Status (TargetNotFound), a non-image URL is 415 (UnsupportedContent, whose message no longer claims the URL 'does not serve HTML').
0.10.1: TikTok Shop search without a keyword, ALDI specials by drop d… …ate; docs wording tiktok shop search keyword optional (category-only search), aldi_au_specials takes category (a Special Buys drop date), capterra search drops page and app store reviews drops lang (the API ignores both there), proxy_country is described as the proxy country in the hand-written methods.
0.10.0: methods for every public plugin endpoint 104 plugin methods (e-commerce, groceries, travel, Google, search engines, real estate, vehicles, jobs, reviews, apps, video/music, SEO), generated from the public OpenAPI spec, plus plugin(path, params) for any endpoint. Unset options are not sent, so the API defaults apply. List values repeat the query key. README: plugin method table; the TikTok and Southwest sections are gone (the methods stay).
0.9.0: BudgetExceeded error for 402 User set budget exceeded The API now answers 402 with the body 'User set budget exceeded' once the account's own monthly budget limit (set in the portal profile) is reached. It gets its own subclass of the 402 billing error, like the other blocks. Claude-Session: https://claude.ai/code/session_01EdVcVzdAE1SK9Ve3u2Acpd
get_parsed: a page with no structured data returns ParsedPage, not an… … error The API now answers parsed_data on a page that rendered but held no structured data with a billed 200: empty data, data_extracted: false, a detail line and the rendered html. get_parsed() (sync and async) returns it as a ParsedPage with the new data_extracted, html and detail attributes instead of raising NoDataExtractedError. The class stays exported, deprecated, and still maps a legacy 422 body. Version 0.8.0. Claude-Session: https://claude.ai/code/session_01P8nWucMM6mZCPYuZH4zXWv
PreviousNext