1 // Copyright 2013 The Chromium Authors. All rights reserved.
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
5 #ifndef COMPONENTS_DOM_DISTILLER_CORE_DISTILLER_H_
6 #define COMPONENTS_DOM_DISTILLER_CORE_DISTILLER_H_
11 #include "base/callback.h"
12 #include "base/containers/hash_tables.h"
13 #include "base/memory/ref_counted.h"
14 #include "base/memory/scoped_ptr.h"
15 #include "base/memory/scoped_vector.h"
16 #include "base/memory/weak_ptr.h"
17 #include "components/dom_distiller/core/article_distillation_update.h"
18 #include "components/dom_distiller/core/distiller_page.h"
19 #include "components/dom_distiller/core/distiller_url_fetcher.h"
20 #include "components/dom_distiller/core/proto/distilled_article.pb.h"
21 #include "net/url_request/url_request_context_getter.h"
24 namespace dom_distiller {
30 typedef base::Callback<void(scoped_ptr<DistilledArticleProto>)>
31 DistillationFinishedCallback;
32 typedef base::Callback<void(const ArticleDistillationUpdate&)>
33 DistillationUpdateCallback;
35 virtual ~Distiller() {}
37 // Distills a page, and asynchronously returns the article HTML to the
38 // supplied |finished_cb| callback. |update_cb| is invoked whenever article
39 // under distillation is updated with more data.
40 // E.g. when distilling a 2 page article, |update_cb| may be invoked each time
41 // a distilled page is added and |finished_cb| will be invoked once
42 // distillation is completed.
43 virtual void DistillPage(const GURL& url,
44 scoped_ptr<DistillerPage> distiller_page,
45 const DistillationFinishedCallback& finished_cb,
46 const DistillationUpdateCallback& update_cb) = 0;
49 class DistillerFactory {
51 virtual scoped_ptr<Distiller> CreateDistiller() = 0;
52 virtual ~DistillerFactory() {}
55 // Factory for creating a Distiller.
56 class DistillerFactoryImpl : public DistillerFactory {
59 scoped_ptr<DistillerURLFetcherFactory> distiller_url_fetcher_factory,
60 const dom_distiller::proto::DomDistillerOptions& dom_distiller_options);
61 virtual ~DistillerFactoryImpl();
62 virtual scoped_ptr<Distiller> CreateDistiller() OVERRIDE;
65 scoped_ptr<DistillerURLFetcherFactory> distiller_url_fetcher_factory_;
66 dom_distiller::proto::DomDistillerOptions dom_distiller_options_;
69 // Distills a article from a page and associated pages.
70 class DistillerImpl : public Distiller {
73 const DistillerURLFetcherFactory& distiller_url_fetcher_factory,
74 const dom_distiller::proto::DomDistillerOptions& dom_distiller_options);
75 virtual ~DistillerImpl();
77 virtual void DistillPage(
79 scoped_ptr<DistillerPage> distiller_page,
80 const DistillationFinishedCallback& finished_cb,
81 const DistillationUpdateCallback& update_cb) OVERRIDE;
83 void SetMaxNumPagesInArticle(size_t max_num_pages);
86 // In case of multiple pages, the Distiller maintains state of multiple pages
87 // as page numbers relative to the page number where distillation started.
88 // E.g. if distillation starts at page 2 for a 3 page article. The relative
89 // page numbers assigned to pages will be [-1,0,1].
91 // Class representing the state of a page under distillation.
92 struct DistilledPageData {
94 virtual ~DistilledPageData();
95 // Relative page number of the page.
97 ScopedVector<DistillerURLFetcher> image_fetchers_;
98 scoped_refptr<base::RefCountedData<DistilledPageProto> >
102 DISALLOW_COPY_AND_ASSIGN(DistilledPageData);
105 void OnFetchImageDone(int page_num,
106 DistillerURLFetcher* url_fetcher,
107 const std::string& id,
108 const std::string& response);
110 void OnPageDistillationFinished(
112 const GURL& page_url,
113 scoped_ptr<proto::DomDistillerResult> distilled_page,
114 bool distillation_successful);
116 virtual void FetchImage(int page_num,
117 const std::string& image_id,
118 const std::string& item);
120 // Distills the next page.
121 void DistillNextPage();
123 // Adds the |url| to |pages_to_be_distilled| if |page_num| is a valid relative
124 // page number and |url| is valid. Ignores duplicate pages and urls.
125 void AddToDistillationQueue(int page_num, const GURL& url);
127 // Check if |page_num| is a valid relative page number, i.e. page with
128 // |page_num| is either under distillation or has already completed
130 bool IsPageNumberInUse(int page_num) const;
132 bool AreAllPagesFinished() const;
134 // Total number of pages in the article that the distiller knows of, this
135 // includes pages that are pending distillation.
136 size_t TotalPageCount() const;
138 // Runs |finished_cb_| if all distillation callbacks and image fetches are
140 void RunDistillerCallbackIfDone();
142 // Checks if page |distilled_page_data| has finished distillation, including
143 // all image fetches.
144 void AddPageIfDone(int page_num);
146 DistilledPageData* GetPageAtIndex(size_t index) const;
148 // Create an ArticleDistillationUpdate for the current distillation
150 const ArticleDistillationUpdate CreateDistillationUpdate() const;
152 const DistillerURLFetcherFactory& distiller_url_fetcher_factory_;
153 scoped_ptr<DistillerPage> distiller_page_;
155 dom_distiller::proto::DomDistillerOptions dom_distiller_options_;
156 DistillationFinishedCallback finished_cb_;
157 DistillationUpdateCallback update_cb_;
159 // Set of pages that are under distillation or have finished distillation.
160 // |started_pages_index_| and |finished_pages_index_| maintains the mapping
161 // from page number to the indices in |pages_|.
162 ScopedVector<DistilledPageData> pages_;
164 // Maps page numbers of finished pages to the indices in |pages_|.
165 std::map<int, size_t> finished_pages_index_;
167 // Maps page numbers of pages under distillation to the indices in |pages_|.
168 // If a page is |started_pages_| that means it is still waiting for an action
169 // (distillation or image fetch) to finish.
170 base::hash_map<int, size_t> started_pages_index_;
172 // The list of pages that are still waiting for distillation to start.
173 // This is a map, to make distiller prefer distilling lower page numbers
175 std::map<int, GURL> waiting_pages_;
177 // Set to keep track of which urls are already seen by the distiller. Used to
178 // prevent distiller from distilling the same url twice.
179 base::hash_set<std::string> seen_urls_;
181 size_t max_pages_in_article_;
183 bool destruction_allowed_;
185 base::WeakPtrFactory<DistillerImpl> weak_factory_;
187 DISALLOW_COPY_AND_ASSIGN(DistillerImpl);
190 } // namespace dom_distiller
192 #endif // COMPONENTS_DOM_DISTILLER_CORE_DISTILLER_H_