Archiver now saves redirects

This commit is contained in:
Cadence Ember 2023-12-12 11:10:47 +13:00
parent 27c9680f5b
commit a57445abcb
8 changed files with 197 additions and 121 deletions

View file

@ -35,9 +35,12 @@
(output-lines? #t)]))
(define (update-width)
(when (output-progress?)
(case (system-type 'os)
[(linux)
(with-charterm
(call-with-values (λ () (charterm-screen-size))
(λ (cols rows) (set! width cols))))))
(λ (cols rows) (set! width cols))))]
[else 100])))
(update-width)
;; check
(when (or (not wikiname) (equal? wikiname ""))
@ -56,8 +59,8 @@
(define real-width (min (string-length basename) rest))
(define spare-width (- rest real-width))
(define name-display (substring basename 0 real-width))
(define whitespace (make-string spare-width #\ ))
(printf "~a~a~a\r" prefix name-display whitespace)]))
(printf "\e[2K\r~a~a" prefix name-display)
(flush-output)]))
;; download all stages
(for ([stage all-stages]
[i (in-naturals 1)])

View file

@ -48,7 +48,8 @@
((query-exec slc* "alter table wiki add column sitename TEXT")
(query-exec slc* "alter table wiki add column basepage TEXT")
(query-exec slc* "alter table wiki add column license_text TEXT")
(query-exec slc* "alter table wiki add column license_url TEXT"))))
(query-exec slc* "alter table wiki add column license_url TEXT"))
((query-exec slc* "alter table page add column redirect"))))
(let do-migrate-step ()
(when (database-version . < . (length migrations))

View file

@ -1,5 +1,6 @@
#lang racket/base
(require racket/file
racket/format
racket/function
racket/list
racket/path
@ -43,10 +44,18 @@
wikiname
(params->query '(("action" . "query")
("meta" . "siteinfo")
("siprop" . "general|rightsinfo|statistics")
("siprop" . "general|rightsinfo|statistics|namespaces")
("format" . "json")
("formatversion" . "2")))))
(define data (response-json (get dest-url)))
(define content-nss
(sort
(for/list ([(k v) (in-hash (jp "/query/namespaces" data))]
#:do [(define id (hash-ref v 'id))]
#:when (and (id . < . 2900) ; exclude maps namespace
(hash-ref v 'content))) ; exclude non-content and talk namespaces
id)
<))
(define exists? (query-maybe-value* "select progress from wiki where wikiname = ?" wikiname))
(if (and exists? (not (sql-null? exists?)))
(query-exec* "update wiki set sitename = ?, basepage = ?, license_text = ?, license_url = ? where wikiname = ?"
@ -61,7 +70,8 @@
(second (regexp-match #rx"/wiki/(.*)" (jp "/query/general/base" data)))
(jp "/query/rightsinfo/text" data)
(jp "/query/rightsinfo/url" data)))
(jp "/query/statistics/articles" data))
(values (jp "/query/statistics/articles" data)
content-nss))
(define (check-style-for-images wikiname path)
@ -131,48 +141,57 @@
;; done yet?
(unless (and (real? wiki-progress) (wiki-progress . >= . 1))
;; Count total pages
(define num-pages (insert-wiki-entry wikiname))
(define-values (num-pages namespaces) (insert-wiki-entry wikiname))
;; Download the entire index of pages
(define basenames
(let loop ([path-with-namefrom "/wiki/Local_Sitemap"]
[basenames-previous-pages null])
;; Download the current index page
(define url (format "https://~a.fandom.com~a" wikiname path-with-namefrom))
(define r (get url))
;; Metadata from this page (the link to the next page)
(define page (html->xexp (bytes->string/utf-8 (response-body r))))
(define link-namefrom
((query-selector (λ (t a c x) (and (eq? t 'a)
(pair? x)
(string-contains? (car x) "Next page")
(let ([href (get-attribute 'href a)] )
(and href (string-contains? href "/wiki/Local_Sitemap")))))
page #:include-text? #t)))
(for*/fold ([total 0])
([namespace namespaces]
[redir-filter '("nonredirects" "redirects")])
(let loop ([apcontinue ""]
[basenames null])
(cond
[apcontinue
(define url (format "https://~a.fandom.com/api.php?~a"
wikiname
(params->query `(("action" . "query")
("list" . "allpages")
("apnamespace" . ,(~a namespace))
("apfilterredir" . ,redir-filter)
("aplimit" . "500")
("apcontinue" . ,apcontinue)
("format" . "json")
("formatversion" . "2")))))
;; Download the current listing page
(define res (get url))
(define json (response-json res))
;; Content from this page
(define basenames-this-page
(for/list ([link (in-producer
(query-selector
(λ (t a c) (eq? t 'a))
((query-selector (λ (t a c) (has-class? "mw-allpages-chunk" a)) page)))
#f)])
(local-encoded-url->basename (get-attribute 'href (bits->attributes link)))))
;; Call the progress callback
(define all-basenames (append basenames-previous-pages basenames-this-page))
(callback (length all-basenames) num-pages (last all-basenames))
;; Recurse to download from the next page
(if link-namefrom
(loop (get-attribute 'href (bits->attributes link-namefrom)) all-basenames)
all-basenames)))
;; Save those pages into the database
(define current-basenames
(for/list ([page (jp "/query/allpages" json)])
(title->basename (jp "/title" page))))
(when ((length current-basenames) . > . 0)
;; Report
(if (equal? redir-filter "nonredirects")
(callback (+ (length basenames) (length current-basenames) total) num-pages (last current-basenames))
(callback total num-pages (last current-basenames))))
;; Loop
(loop (jp "/continue/apcontinue" json #f) (append basenames current-basenames))]
[else
;; All done with this (loop)! Save those pages into the database
;; SQLite can have a maximum of 32766 parameters in a single query
(for ([slice (in-slice 32760 basenames)])
(define query-template (string-join (make-list (length slice) "(?1, ?, 0)") ", " #:before-first "insert or ignore into page (wikiname, basename, progress) values "))
(begin0
;; next for*/fold
(if (equal? redir-filter "nonredirects")
(+ (length basenames) total)
total) ; redirects don't count for the site statistics total
(call-with-transaction
(get-slc)
(λ ()
(apply query-exec* query-template wikiname slice)
(for ([slice (in-slice 32760 basenames)])
(define query-template
(string-join #:before-first "insert or ignore into page (wikiname, redirect, basename, progress) values "
(make-list (length slice) "(?1, ?2, ?, 0)") ", "))
(apply query-exec* query-template wikiname (if (equal? redir-filter "redirects") 1 sql-null) slice)))))])))
;; Record that we have the complete list of pages
(query-exec* "update wiki set progress = 1 where wikiname = ?" wikiname))))))
(query-exec* "update wiki set progress = 1 where wikiname = ?" wikiname)))
;; 2. Download each page via API and:
@ -183,7 +202,7 @@
(define save-dir (build-path archive-root wikiname))
(make-directory* save-dir)
;; gather list of basenames to download (that aren't yet complete)
(define basenames (query-list* "select basename from page where wikiname = ? and progress < ?"
(define basenames (query-list* "select basename from page where wikiname = ? and progress < ? and redirect is null"
wikiname max-page-progress))
;; counter of complete/incomplete basenames
(define already-done-count
@ -222,10 +241,41 @@
(query-exec* "update page set progress = 1 where wikiname = ? and basename = ?"
wikiname basename)
(callback i total-count basename))
;; save redirects as well
(save-redirects wikiname callback (+ already-done-count (length basenames)) total-count)
;; saved all pages, register that fact in the database
(query-exec* "update wiki set progress = 2 where wikiname = ?" wikiname))
;; 2.5. Download each redirect-target via API and save mapping in database
(define (save-redirects wikiname callback already-done-count total-count)
(define basenames (query-list* "select basename from page where wikiname = ? and progress < ? and redirect = 1"
wikiname max-page-progress))
;; loop through basenames, in slices of 50 (MediaWiki API max per request), and download
(for ([basename basenames]
[i (in-naturals (add1 already-done-count))])
(define dest-url
(format "https://~a.fandom.com/api.php?~a"
wikiname
(params->query `(("action" . "query")
("prop" . "links")
("titles" . ,(basename->name-for-query basename))
("format" . "json")
("formatversion" . "2")))))
(define res (get dest-url))
(define json (response-json res))
(define dest-title (jp "/query/pages/0/links/0/title" json #f))
(callback i total-count basename)
(cond
[dest-title
;; store it
(define dest-basename (title->basename dest-title))
(query-exec* "update page set progress = 1, redirect = ? where wikiname = ? and basename = ?" dest-basename wikiname basename)]
[else
;; the page just doesn't exist
(query-exec* "delete from page where wikiname = ? and basename = ?" wikiname basename)])))
;; 3. Download CSS and:
;; * Save CSS to file
;; * Record style images to database

View file

@ -7,6 +7,7 @@
local-encoded-url->segments
url-segments->basename
local-encoded-url->basename
title->basename
basename->name-for-query
url-segments->guess-title)
@ -21,6 +22,12 @@
(define (local-encoded-url->basename str) ; '("wiki" "Page_title"), no extension or dir prefix
(url-segments->basename (local-encoded-url->segments str)))
(define (title->basename title) ; "Page title/Strategies" -> "Page_title#Strategies" filename encoded, no extension or dir prefi
(define elements (string-split (string-replace title " " "_") "/"))
(define extra-encoded (map (λ (s) (bytes->string/latin-1 (percent-encode s filename-set #f))) elements))
(define basic-filename (string-join extra-encoded "#"))
basic-filename)
(define (basename->name-for-query str)
(uri-decode (regexp-replace* #rx"#" str "/")))

View file

@ -377,7 +377,7 @@
'("runescape") 'default
'RuneScape
"RuneScape Wiki"
"https://runescape.wiki/"
"https://runescape.wiki/w/Main_Page"
"https://runescape.wiki/images/Wiki.png"
(λ (props)
`((p "The RuneScape Wiki was founded on April 8, 2005. In October 2018, the wiki left Fandom (then Wikia), citing their apathy towards the wiki and excessive advertisements."))))
@ -386,7 +386,7 @@
'("oldschoolrunescape") 'default
'RuneScape
"Old School RuneScape Wiki"
"https://oldschool.runescape.wiki/"
"https://oldschool.runescape.wiki/w/Main_Page"
"https://oldschool.runescape.wiki/images/Wiki.png"
(λ (props)
`((p "The Old School RuneScape Wiki was founded on February 14, 2013. In October 2018, the RuneScape Wiki left Fandom (then Wikia), citing their apathy towards the wiki and excessive advertisements, with the Old School RuneScape Wiki following suit."))))
@ -395,7 +395,7 @@
'("runescapeclassic") 'default
'RuneScape
"RuneScape Classic Wiki"
"https://classic.runescape.wiki/"
"https://classic.runescape.wiki/w/Main_Page"
"https://classic.runescape.wiki/images/Wiki.png"
(λ (props)
`((p "The Old School RuneScape Wiki was founded on April 19, 2009. In October 2018, the RuneScape Wiki left Fandom (then Wikia), citing their apathy towards the wiki and excessive advertisements, with the RuneScape Classic Wiki following suit."))))

View file

@ -19,7 +19,7 @@
[(not wikiname)
(response/output
#:code 400
#:mime-type "text/plain"
#:mime-type #"text/plain"
(λ (out)
(displayln "Requires wikiname and q parameters." out)))]
[(or (not q) (equal? q ""))

View file

@ -53,7 +53,8 @@
;; grab the part after ?q= which is the search terms
(define query (dict-ref params 'q #f))
;; figure out which search provider we're going to use
(define search-provider (hash-ref search-providers (config-get 'feature_offline::search)))
(define search-provider (hash-ref search-providers (config-get 'feature_offline::search)
(λ () (error 'search-provider "unknown search provider configured"))))
;; external special:search url to link at the bottom of the page as the upstream source
(define external-search-url

View file

@ -20,6 +20,7 @@
web-server/dispatchers/dispatch
; my libs
"application-globals.rkt"
"../archiver/archiver-database.rkt"
"config.rkt"
"data.rkt"
"log.rkt"
@ -40,6 +41,9 @@
(define path-archive (anytime-path ".." "storage/archive"))
(when (config-true? 'feature_offline::only)
(void (get-slc)))
(define (page-wiki-offline req)
(response-handler
(define wikiname (path/param-path (first (url-path (request-uri req)))))
@ -64,39 +68,9 @@
[else (error 'archive-format "unknown archive format configured")]))
(define fs-path (build-path path-archive wikiname (format (car archive-format) maybe-hashed-basename)))
(define source-url (format "https://~a.fandom.com/wiki/~a" wikiname (basename->name-for-query basename)))
(cond
[(not (file-exists? fs-path))
(unless (config-true? 'feature_offline::only)
(next-dispatcher))
(define mirror-path (url->string (request-uri req)))
(define body
(generate-wiki-page
`(div (@ (class "unsaved-page"))
(style ".unsaved-page a { text-decoration: underline !important }")
(p "breezewiki.com doesn't have this page saved.")
(p "You can see this page by visiting a BreezeWiki mirror:")
(ul
(li (a (@ (href ,(format "https://antifandom.com~a" mirror-path))) "View on antifandom.com"))
(li (a (@ (href ,(format "https://bw.artemislena.eu~a" mirror-path))) "View on artemislena.eu"))
(li (a (@ (href ,source-url)) "or, you can see the original page on Fandom (ugh)")))
(p "If you'd like " ,wikiname ".fandom.com to be added to breezewiki.com, " (a (@ (href "https://lists.sr.ht/~cadence/breezewiki-requests")) "let me know about it!")))
#:req req
#:source-url source-url
#:wikiname wikiname
#:title (url-segments->guess-title segments)
#:online-styles #f
#:siteinfo (siteinfo-fetch wikiname)
))
(when (config-true? 'debug)
; used for its side effects
; convert to string with error checking, error will be raised if xexp is invalid
(xexp->html body))
(response/output
#:code 200
#:headers always-headers
(λ (out)
(write-html body out)))]
[#t
(cond/var
[(file-exists? fs-path)
(when (config-true? 'debug)
(printf "using offline mode for ~v~n" fs-path))
(response-handler
@ -144,4 +118,44 @@
#:code 200
#:headers headers
(λ (out)
(write-html body out))))])))
(write-html body out))))]
;; page not found on disk, perhaps it's a redirect? redirects are stored in the database
(var target (query-maybe-value* "select redirect from page where wikiname = ? and basename = ?" wikiname basename))
[target
(generate-redirect (basename->name-for-query target))]
;; breezewiki doesn't have the page archived, see if we can make a network request for it
[(not (config-true? 'feature_offline::only))
(next-dispatcher)]
;; no possible way to provide the page
[else
(define mirror-path (url->string (request-uri req)))
(define body
(generate-wiki-page
`(div (@ (class "unsaved-page"))
(style ".unsaved-page a { text-decoration: underline !important }")
(p "breezewiki.com doesn't have this page saved.")
(p "You can see this page by visiting a BreezeWiki mirror:")
(ul
(li (a (@ (href ,(format "https://antifandom.com~a" mirror-path))) "View on antifandom.com"))
(li (a (@ (href ,(format "https://bw.artemislena.eu~a" mirror-path))) "View on artemislena.eu"))
(li (a (@ (href ,source-url)) "or, you can see the original page on Fandom (ugh)")))
(p "If you'd like " ,wikiname ".fandom.com to be added to breezewiki.com, " (a (@ (href "https://lists.sr.ht/~cadence/breezewiki-requests")) "let me know about it!")))
#:req req
#:source-url source-url
#:wikiname wikiname
#:title (url-segments->guess-title segments)
#:online-styles #f
#:siteinfo (siteinfo-fetch wikiname)
))
(when (config-true? 'debug)
; used for its side effects
; convert to string with error checking, error will be raised if xexp is invalid
(xexp->html body))
(response/output
#:code 200
#:headers always-headers
(λ (out)
(write-html body out)))])))