5 Before results appear in the UrlbarView, they are fetched from providers.
7 Each `UrlbarProvider <https://firefox-source-docs.mozilla.org/browser/urlbar/overview.html#urlbarprovider>`_
8 implements its own internal ranking and returns sorted results.
10 Externally all the results are ranked by the `UrlbarMuxer <https://searchfox.org/mozilla-central/source/browser/components/urlbar/UrlbarMuxerUnifiedComplete.sys.mjs>`_
11 according to an hardcoded list of groups and sub-grups.
13 .. NOTE:: Preferences can influence the groups order, for example by putting
14 Firefox Suggest before Search Suggestions.
16 The Places provider, responsible to return history and bookmark results, uses
17 an internal ranking algorithm called Frecency.
19 Frecency implementation
20 =======================
22 Frecency is a term derived from `frequency` and `recency`, its scope is to provide a
23 ranking algorithm that gives importance both to how often a page is accessed and
24 when it was last visited.
25 Additionally, it accounts for the type of each visit through a bonus system.
27 To account for `recency`, a bucketing system is implemented.
28 If a page has been visited later than the bucket cutoff, it gets the weight
29 associated with that bucket:
31 - Up to 4 days old - weight 100 - ``places.frecency.firstBucketCutoff/Weight``
32 - Up to 14 days old - weight 70 - ``places.frecency.secondBucketCutoff/Weight``
33 - Up to 31 days old - weight 50 - ``places.frecency.thirdBucketCutoff/Weight``
34 - Up to 90 days old - weight 30 - ``places.frecency.fourthBucketCutoff/Weight``
35 - Anything else - weight 10 - ``places.frecency.defaultBucketWeight``
37 To account for `frequency`, the total number of visits to a page is used to
38 calculate the final score.
40 The type of each visit is taken into account using specific bonuses:
43 Any unknown type gets a default bonus. This is expected to be unused.
44 Pref ``places.frecency.defaultVisitBonus`` current value: 0.
46 Used for embedded/framed visits not due to user actions. These visits today
47 are stored in memory and never participate to frecency calculation.
48 Thus this is currently unused.
49 Pref ``places.frecency.embedVisitBonus`` current value: 0.
51 Used for cross-frame visits due to user action.
52 Pref ``places.frecency.framedLinkVisitBonus`` current value: 0.
54 Used for download visits. It’s important to support link coloring for these
55 visits, but they are not necessarily useful address bar results (the Downloads
56 view can do a better job with these), so their frecency can be low.
57 Pref ``places.frecency.downloadVisitBonus`` current value: 0.
59 Used for reload visits (refresh same page). Low because it should not be
60 possible to influence frecency by multiple reloads.
61 Pref ``places.frecency.reloadVisitBonus`` current value: 0.
63 Used when the page redirects to another one.
64 It’s a low value because we give more importance to the final destination,
65 that is what the user actually visits, especially for permanent redirects.
66 Pref ``places.frecency.redirectSourceVisitBonus`` current value: 25.
68 Used for visits resulting from a temporary redirect (HTTP 307).
69 Pref ``places.frecency.tempRedirectVisitBonus`` current value: 40.
71 Used for visits resulting from a permanent redirect (HTTP 301). This is the
72 new supposed destination for a url, thus the bonus is higher than temporary.
73 In this case it may be advisable to just pick the bonus for the source visit.
74 Pref ``places.frecency.permRedirectVisitBonus`` current value: 50.
76 Used for visits generated from bookmark views.
77 Pref ``places.frecency.bookmarkVisitBonus`` current value: 75.
79 Used for normal visits, for example when clicking on a link.
80 Pref ``places.frecency.linkVisitBonus`` current value: 100.
82 Intended to be used for pages typed by the user, in reality it is used when
83 the user picks a url from the UI (history views or the Address Bar).
84 Pref ``places.frecency.typedVisitBonus`` current value: 2000.
86 The above bonuses are applied to visits, in addition to that there are also a
87 few bonuses applied in case a page is not visited at all, both of these bonuses
88 can be applied at the same time:
90 Unvisited bookmarked page
91 Used for pages that are bookmarked but unvisited.
92 Pref ``places.frecency.unvisitedBookmarkBonus`` current value: 140.
94 Used for pages that were typed and now are bookmarked (otherwise they would
96 Pref ``places.frecency.unvisitedTypedBonus`` current value: 200.
98 Two special frecency values are also defined:
100 - ``-1`` represents a just inserted entry in the database, whose score has not
102 - ``0`` represents an entry for which a new value should not be calculated,
103 because it has a poor user value (e.g. place: queries) among search results.
105 Finally, because calculating a score from all of the visits every time a new
106 visit is added would be expensive, only a sample of the last 10
107 (pref ``places.frecency.numVisits``) visits is used.
109 How frecency for a page is calculated
110 -------------------------------------
114 :caption: Frecency calculation flow
119 a1[Get last 10 visit]
120 a2[bonus = unvisited_bonus + bookmarked + typed]
123 end1["frecency = age_bucket_weight * (bonus / 100)"]
124 a4[Sum points of all sampled visits]
127 end3["Frecency = visit_count * (points / sample_size)"]
128 subgraph sub [Per each visit]
129 sub0[bonus = visit_type_bonus]
131 sub2[add bookmark bonus]
132 sub3["score = age_bucket_weight * (bonus / 100)"]
150 1. If the page is visited, get a sample of ``NUM_VISITS`` most recent visits.
151 2. For each visit get a transition bonus, depending on the visit type.
152 3. If the page is bookmarked, add to the bonus an additional bookmark bonus.
153 4. If the bonus is positive, get a bucket weight depending on the visit date.
154 5. Calculate points for the visit as ``age_bucket_weight * (bonus / 100)``.
155 6. Sum points for all the sampled visits.
156 7. If the points sum is zero, return a ``-1`` frecency, it will still appear in the UI.
157 Otherwise, frecency is ``visitCount * points / NUM_VISITS``.
158 8. If the page is unvisited and not bookmarked, or it’s a bookmarked place-query,
159 return a ``0`` frecency, to hide it from the UI.
160 9. If it’s bookmarked, add the bookmark bonus.
161 10. If it’s also a typed page, add the typed bonus.
162 11. Frecency is ``age_bucket_weight * (bonus / 100)``
164 When frecency for a page is calculated
165 --------------------------------------
167 Operations that may influence the frecency score are:
173 * Changing the url of a bookmark
175 Frecency is recalculated:
177 * Immediately, when a new visit is added. The user expectation here is that the
178 page appears in search results after being visited. This is also valid for
179 any History API that allows to add visits.
180 * In background on idle times, in any other case. In most cases having a
181 temporarily stale value is not a problem, the main concern would be privacy
182 when removing history of a page, but removing whole history will either
183 completely remove the page or, if it's bookmarked, it will still be relevant.
184 In this case, when a change influencing frecency happens, the ``recalc_frecency``
185 database field for the page is set to ``1``.
187 Recalculation is done by the `PlacesFrecencyRecalculator <https://searchfox.org/mozilla-central/source/toolkit/components/places/PlacesFrecencyRecalculator.sys.mjs>`_ module.
188 The Recalculator is notified when ``PlacesUtils.history.shouldStartFrecencyRecalculation``
189 value changes from false to true, that means there's values to recalculate.
190 A DeferredTask is armed, that will look for a user idle opportunity
191 in the next 5 minutes, otherwise it will run when that time elapses.
192 Once all the outdated values have been recalculated
193 ``PlacesUtils.history.shouldStartFrecencyRecalculation`` is set back to false
194 until the next operation invalidating a frecency.
195 The recalculation task is also armed on the ``idle-daily`` notification.
197 When the task is executed, it recalculates frecency of a chunk of pages. If
198 there are more pages left to recalculate, the task is re-armed. After frecency
199 of a page is recalculated, its ``recalc_frecency`` field is set back to ``0``.
201 Frecency is also decayed daily during the ``idle-daily`` notification, by
202 multiplying all the scores by a decay rate of ``0.975`` (half-life of 28 days).
203 This guarantees entries not receiving new visits or bookmarks lose relevancy.
206 Adaptive Input History
207 ======================
209 Input History (also known as Adaptive History) is a feature that allows to
210 find urls that the user previously picked. To do so, it associates search strings
213 Adaptive results are usually presented before frecency derived results, making
214 them appear as having an infinite frecency.
216 When the user types a given string, and picks a result from the address bar, that
217 relation is stored and increases a use_count field for the given string.
218 The use_count field asymptotically approaches a max of ``10`` (the update is
219 done as ``use_count * .9 + 1``).
221 On querying, all the search strings that start with the input string are matched,
222 a rank is calculated per each page as ``ROUND(MAX(use_count) * (1 + (input = :search_string)), 1)``,
223 so that results perfectly matching the search string appear at the top.
224 Results with the same rank are additionally sorted by descending frecency.
226 On daily idles, when frecency is decayed, also input history gets decayed, in
227 particular the use_count field is multiplied by a decay rate of ``0.975``.
228 After decaying, any entry that has a ``use_count < 0.975^90 (= 0.1)`` is removed,
229 thus entries are removed if unused for 90 days.