pinterest-dl 2.0.5 → 2.1.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d14e37c46d9bf7d3b8851c433a430da6811a040151e1103912b300985569b038
-  data.tar.gz: f19617c2661b7cb970206eabe8a6fb1471dfa9045f09ebf2e0393bfebdc83e5c
+  metadata.gz: 51b7122dcca9ae653e0f63a70665fa8b5dfddd87a9d7e17eeef3462843ed90c1
+  data.tar.gz: e82c0d139c5bf84dfd5e13fa96488d8013636f72ce7e2bdc5bc9a8efc49151cc
 SHA512:
-  metadata.gz: 8ac5b7a9d1525a24dee57a7857cc345d8febb84061cb82b018d507c35c03d04e2d214cc108584dcfb56208d6f05f524a74fa427faa9ffd1b25e34314ab597eae
-  data.tar.gz: 565a098363fbf28418ef5312972bf2e4dec08d8ade89cad70a59c425bc1a25f052c56eb4c335e48b15237097db9487c833384d4ede470acd685cec986ed98e82
+  metadata.gz: 911490799bd9db4405781306091bbe06271413e25b834190d46394ad7b288be5ce3153ba25d9261122dbb4efe0be08188c49583751999d08e87be223a017300e
+  data.tar.gz: f18ffdcc24defdbf6bb392f738ed6b6dc62047ccecacfb90eb90fb52a879b18dd84c7677288d538f8105ecae7e611faec6470a0dfeaedb07f8767d9e822c0e20

data/CHANGELOG.md

@@ -3,7 +3,7 @@
 All notable changes to this project are documented here.
 This project follows [Semantic Versioning](https://semver.org/).
 
-## [2.0.5]
+## [2.1.5]
 
 ### Fixed
 - Pin URLs on country domains (e.g. `de.pinterest.com/pin/...`) didn't download
@@ -17,6 +17,14 @@ This project follows [Semantic Versioning](https://semver.org/).
 - Video pins returned "No video found". The best `.mp4` is picked now (not just
   the first format, which could be an `.m3u8`), story pins are supported, and
   `og:video` meta tags are read in any attribute order.
+- `video` worked only some of the time: Pinterest sometimes serves a lighter
+  page without the video data. Now every `__PWS_*__` data block of the page is
+  read, then the pin is asked from the `PinResource` API (with the
+  `X-Pinterest-PWS-Handler` header Pinterest needs, otherwise it answers 403),
+  and the page is fetched fresh up to 3 times before giving up. The "this is an
+  image" hint only shows when Pinterest really returned an image pin.
+- `board` reads the board id from the page first and only asks the API when it
+  isn't there. Story modules ("Related interests") are no longer treated as pins.
 - `board` never worked: Pinterest needs the real board id, we sent an empty
   one. The id is now looked up first, the session is created once instead of
   once per page, links with `?invite_code=...` are parsed correctly, and
@@ -89,4 +97,4 @@ This project follows [Semantic Versioning](https://semver.org/).
 ## [1.0.3]
 
 - Initial public release: `get_image_url`, `get_video_url`, `search`,
-  `search_json`, `search_images`, `search_videos`.
+  `search_json`, `search_images`, `search_videos`.

data/bin/pinterest-dl

@@ -19,7 +19,6 @@ def usage
       -t, --type TYPE     What to collect in search/board: all, image, video, pin (default all)
       -q, --quality Q     Image quality: originals, 736x, 474x, 236x (default originals)
       -c, --cookies STR   Cookie string for authenticated requests
-      -v, --verbose       Print debug logging
       -h, --help          Show this help
   USAGE
 end
@@ -31,7 +30,6 @@ parser = OptionParser.new do |opts|
   opts.on('-t T', '--type T') { |v| options[:type] = v.to_sym }
   opts.on('-q Q', '--quality Q') { |v| options[:quality] = v.to_sym }
   opts.on('-c STR', '--cookies STR') { |v| options[:cookies] = v }
-  opts.on('-v', '--verbose') { options[:verbose] = true }
   opts.on('-h', '--help') do
     puts usage
     exit 0

data/lib/pinterest_dl/api.rb

@@ -21,20 +21,25 @@ module PinterestDL
     def get_image_url(pin_url, quality: config.quality, client: Client.new(config: config))
       html, pin_id = load_pin(pin_url, client)
       url = extract_pin_image(html, pin_id, quality: quality)
+      url ||= api_image_url(pin_id, client, quality)
       raise NotFoundError, "No image found at #{pin_url}" unless url
 
       url
     end
 
     def get_video_url(pin_url, client: Client.new(config: config))
-      html, pin_id = load_pin(pin_url, client)
-      url = extract_pin_video(html, pin_id)
-      unless url
-        hint = extract_pin_image(html, pin_id, quality: :originals) ? ' (this pin is an image, try the image command)' : ''
-        raise NotFoundError, "No video found at #{pin_url}#{hint}"
+      pin = nil
+      3.times do |attempt|
+        html, pin_id = load_pin(pin_url, client, fresh: attempt.positive?)
+        url = extract_pin_video(html, pin_id)
+        return url if url
+
+        pin = pin_from_api(pin_id, client)
+        return pin[:video_url] if pin && pin[:video_url]
       end
 
-      url
+      hint = pin && pin[:image_url] ? ' (it looks like an image pin, try the image command)' : ''
+      raise NotFoundError, "No video found at #{pin_url}#{hint}"
     end
 
     def get_media(pin_url, quality: config.quality, client: Client.new(config: config))
@@ -79,10 +84,40 @@ module PinterestDL
 
     private
 
-    def load_pin(pin_url, client)
+    def load_pin(pin_url, client, fresh: false)
       validate_pin_url!(pin_url)
-      html = client.get(normalize_pin_url(pin_url))
-      [html, pin_id_from(client.last_url) || pin_id_from(pin_url)]
+      html = client.get(normalize_pin_url(pin_url), use_cache: !fresh)
+      id = pin_id_from(client.last_url) || pin_id_from(pin_url) || html.to_s[%r{pinterest\.[a-z.]+/pin/(\d+)}, 1]
+      [html, id]
+    end
+
+    def pin_from_api(pin_id, client)
+      return nil unless pin_id
+
+      data = { options: { id: pin_id, field_set_key: 'detailed' }, context: {} }.to_json
+      uri = URI('https://www.pinterest.com/resource/PinResource/get/')
+      uri.query = URI.encode_www_form('source_url' => "/pin/#{pin_id}/", 'data' => data)
+      headers = {
+        'Accept' => 'application/json, text/javascript, */*, q=0.01',
+        'X-Requested-With' => 'XMLHttpRequest',
+        'X-Pinterest-PWS-Handler' => 'www/pin/[id].js',
+        'X-Pinterest-Source-Url' => "/pin/#{pin_id}/",
+        'X-CSRFToken' => client.cookie('csrftoken') || 'undefined',
+        'Referer' => "https://www.pinterest.com/pin/#{pin_id}/"
+      }
+
+      pin = JSON.parse(client.get(uri.to_s, headers: headers, use_cache: false)).dig('resource_response', 'data')
+      return nil unless pin.is_a?(Hash)
+
+      { image_url: pin.dig('images', 'orig', 'url'), video_url: Extractors::PinData.first_video_url(pin) }
+    rescue PinterestDL::Error, JSON::ParserError => e
+      logger.debug("PinResource failed for pin #{pin_id}: #{e.message}")
+      nil
+    end
+
+    def api_image_url(pin_id, client, quality)
+      url = pin_from_api(pin_id, client)&.dig(:image_url)
+      url && Extractors::Image.apply_quality(url, quality)
     end
 
     def normalize_pin_url(pin_url)

data/lib/pinterest_dl/board.rb

@@ -14,8 +14,8 @@ module PinterestDL
 
     def call(board_url, limit: 50, pages: 10, type: :all)
       username, slug, section = parse_url(board_url)
-      start_session(username, slug)
-      board_id = find_board_id(username, slug)
+      html = start_session(username, slug)
+      board_id = board_id_from_page(html, slug) || find_board_id(username, slug)
       resource, options = feed_request(board_id, section)
 
       results = []
@@ -57,10 +57,34 @@ module PinterestDL
     end
 
     def start_session(username, slug)
-      @client.raw_get("https://www.pinterest.com/#{username}/#{slug}/")
+      response = @client.raw_get("https://www.pinterest.com/#{username}/#{slug}/")
       @page_url = @client.last_url
       @host = URI(@page_url).host
       @source_path = URI(@page_url).path
+      response.body
+    end
+
+    def board_id_from_page(html, slug)
+      html.to_s.scan(PinterestDL::Extractors::PinData::SCRIPT_PATTERN).each do |(raw)|
+        data = PinterestDL::Extractors::PinData.safe_parse(raw)
+        stack = [data]
+        until stack.empty?
+          node = stack.pop
+          case node
+          when Hash
+            return node['id'].to_s if board_node?(node, slug)
+
+            stack.concat(node.values)
+          when Array
+            stack.concat(node)
+          end
+        end
+      end
+      nil
+    end
+
+    def board_node?(node, slug)
+      node['type'] == 'board' && node['id'].to_s.match?(/\A\d+\z/) && node['url'].to_s.include?("/#{slug}/")
     end
 
     def find_board_id(username, slug)
@@ -130,6 +154,7 @@ module PinterestDL
 
     def build_pin(raw)
       return nil unless raw.is_a?(Hash)
+      return nil if raw['type'] && raw['type'] != 'pin'
 
       id = raw['id']
       return nil unless id

data/lib/pinterest_dl/extractors/pin_data.rb

@@ -5,17 +5,26 @@ require 'json'
 module PinterestDL
   module Extractors
     module PinData
-      SCRIPT_PATTERN = %r{<script[^>]*id=["']__PWS_DATA__["'][^>]*>(.*?)</script>}m.freeze
+      SCRIPT_PATTERN = %r{<script[^>]*id=["']__PWS_[A-Z_]+__["'][^>]*>(.*?)</script>}m.freeze
 
       module_function
 
       def extract(html, pin_id = nil)
-        data = parse_embedded_json(html)
-        return nil unless data
+        found = nil
+        html.to_s.scan(SCRIPT_PATTERN).each do |(raw)|
+          data = safe_parse(raw)
+          pin = data && find_pin_node(data, pin_id)
+          next unless pin
+
+          result = build_result(pin)
+          return result if result[:video_url]
 
-        pin = find_pin_node(data, pin_id)
-        return nil unless pin
+          found ||= result
+        end
+        found
+      end
 
+      def build_result(pin)
         {
           image_url: pin.dig('images', 'orig', 'url'),
           video_url: first_video_url(pin),
@@ -76,7 +85,15 @@ module PinterestDL
         return best['url'] if best
 
         hls = formats.find { |v| v['url'].include?('.m3u8') }
-        hls && hls['url'].sub('/hls/', '/720p/').sub(/\.m3u8.*\z/, '.mp4')
+        hls && hls_to_mp4(hls['url'])
+      end
+
+      def hls_to_mp4(url)
+        if url.include?('/iht/hls/')
+          url.sub('/hls/', '/expMp4/').sub(/\.m3u8.*\z/, '_720w.mp4')
+        else
+          url.sub('/hls/', '/720p/').sub(/\.m3u8.*\z/, '.mp4')
+        end
       end
 
       def deep_find(node, key)

data/lib/pinterest_dl/search.rb

@@ -49,9 +49,6 @@ module PinterestDL
 
       parse_search_response(json)
     end
-
-    # Pinterest's search resource requires a session cookie + CSRF token
-    # obtained by first loading the human-facing search page.
     def establish_session(query)
       source_path = "/search/pins/?q=#{URI.encode_www_form_component(query)}"
       home_uri = URI("https://www.pinterest.com#{source_path}")

data/lib/pinterest_dl/version.rb

@@ -2,5 +2,5 @@
 
 # Monji024
 module PinterestDL
-  VERSION = '2.0.5'
+  VERSION = '2.1.5'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: pinterest-dl
 version: !ruby/object:Gem::Version
-  version: 2.0.5
+  version: 2.1.5
 platform: ruby
 authors:
 - Monji
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-10-12 00:00:00.000000000 Z
+date: 2026-10-14 00:00:00.000000000 Z
 dependencies: []
 description: A Ruby toolkit for getting direct image and video URLs from Pinterest
   pins, searching pins, fetching board contents, and downloading media to disk — as