X-Git-Url: https://git.xandkar.net/?a=blobdiff_plain;f=tt.rkt;h=e93c26f4a20e67426ec608c863204e638dead67f;hb=HEAD;hp=58419dc5ce6d8ed778c494939118cde03d6a9b55;hpb=5b27fc74ad4fc719fd6dcffefe776cdc06aadf9d;p=tt.git diff --git a/tt.rkt b/tt.rkt index 58419dc..e93c26f 100644 --- a/tt.rkt +++ b/tt.rkt @@ -27,16 +27,24 @@ (∀ (α β) (U (cons 'ok α) (cons 'error β)))) +(define-type Download-Result + (Result (U 'skipped-cached 'downloaded-new) + (U 'timeout + (Pair 'unsupported-url-scheme String) + (Pair 'http-not-ok Positive-Integer) + (Pair 'net-error Any) + (Pair 'other Any)))) + (struct Hist ([freq : Nonnegative-Integer] [last : Nonnegative-Integer]) #:transparent) -(define-type Nick-Hist +(define-type Url-Nick-Hist (Immutable-HashTable Url (Immutable-HashTable (Option String) Hist))) (struct User - ([uri : Url] + ([url : Url] [nick : (Option String)])) (struct User-Agent @@ -56,8 +64,8 @@ (struct Peer ([nick : (Option String)] - [uri : Url] - [uri-str : String] + [url : Url] + [url-str : String] [comment : (Option String)]) #:transparent) @@ -131,12 +139,12 @@ (: peers-equal? (-> Peer Peer Boolean)) (define (peers-equal? p1 p2) - (equal? (Peer-uri-str p1) - (Peer-uri-str p2))) + (equal? (Peer-url-str p1) + (Peer-url-str p2))) (: peer-hash (-> Peer Fixnum)) (define (peer-hash p) - (equal-hash-code (Peer-uri-str p))) + (equal-hash-code (Peer-url-str p))) (define-custom-set-types peers #:elem? Peer? @@ -158,29 +166,19 @@ (: peers-merge (-> (Listof Peer) * (Listof Peer))) (define (peers-merge . peer-sets) - (define groups - (foldl - (λ (p groups) - (hash-update groups (Peer-uri-str p) (λ (group) (cons p group)) '())) - (hash) - (append* peer-sets))) - (define (merge peers) + (define (merge-2 p1 p2) + (match* (p1 p2) + [((Peer n1 _ _ _) (Peer n2 _ _ _)) #:when (and n1 n2) p1] ; TODO compare which is more-common? + [((Peer #f _ _ _) (Peer #f _ _ _)) p1] ; TODO update with most-common nick? + [((Peer n1 _ _ _) (Peer #f _ _ _)) p1] + [((Peer #f _ _ _) (Peer n2 _ _ _)) p2])) + (: merge-n (-> (Listof Peer) Peer)) + (define (merge-n peers) (match peers ['() (raise 'impossible)] [(list p) p] - [(list* p1 p2 ps) - (let* ([n1 (Peer-nick p1)] - [n2 (Peer-nick p2)] - [p (cond - ; TODO Try to pick from nicks db: preferred, otherwise seen - [(and (not n1) (not n2)) p1] ; TODO update with most-common nick - [(and n1 n2 ) p1] ; TODO compare which is more-common - [(and n1 (not n2)) p1] - [(and (not n1) n2) p2] - [else - (raise 'impossible)])]) - (merge (cons p ps)))])) - (sort (map merge (hash-values groups)) + [(list* p1 p2 ps) (merge-n (cons (merge-2 p1 p2) ps))])) + (sort (map merge-n (group-by Peer-url-str (append* peer-sets))) (match-lambda** [((Peer _ _ u1 _) (Peer _ _ u2 _)) (string\033[0m \033[0;~am~a\033[0m~n" (parameterize ([date-display-format 'iso-8601]) @@ -270,7 +268,7 @@ ([date-display-format 'rfc2822]) (date->string (seconds->date (Msg-ts-epoch msg)) #t)) (Msg-ts-orig msg) - nick uri color text))]))))) + nick url color text))]))))) (: rfc3339->epoch (-> String (Option Nonnegative-Integer))) (define rfc3339->epoch @@ -326,8 +324,8 @@ (let ([mentions (filter-map (λ (m) (match (regexp-match #px"@<([^>]+)>" m) - [(list _wholething nick-uri) - (str->peer nick-uri)])) + [(list _wholething nick-url) + (str->peer nick-url)])) (regexp-match* #px"@<[^\\s]+([\\s]+)?[^>]+>" text))]) (Msg ts-epoch ts-orig from text mentions)) (begin @@ -347,7 +345,7 @@ (string-append d h m))] [tzs (list* "" "Z" tzs)]) (for* ([n '("fake-nick")] - [u '("http://fake-uri")] + [u '("http://fake-url")] [p (list (Peer n (string->url u) u #f))] [s '("" ":10")] [f '("" ".1337")] @@ -368,8 +366,8 @@ [tab " "] [text "Lorem ipsum"] [nick "foo"] - [uri "http://bar/"] - [peer (Peer nick (string->url uri) uri #f)] + [url "http://bar/"] + [peer (Peer nick (string->url url) url #f)] [actual (str->msg peer (string-append ts tab text))] [expected (Msg 1605756129 ts peer text '())]) (check-equal? @@ -385,13 +383,13 @@ (Peer-nick (Msg-from expected)) "str->msg nick") (check-equal? - (Peer-uri (Msg-from actual)) - (Peer-uri (Msg-from expected)) - "str->msg uri") + (Peer-url (Msg-from actual)) + (Peer-url (Msg-from expected)) + "str->msg url") (check-equal? - (Peer-uri-str (Msg-from actual)) - (Peer-uri-str (Msg-from expected)) - "str->msg uri-str") + (Peer-url-str (Msg-from actual)) + (Peer-url-str (Msg-from expected)) + "str->msg url-str") (check-equal? (Msg-text actual) (Msg-text expected) @@ -417,37 +415,37 @@ (define cache-object-dir (build-path cache-dir "objects")) (: url->cache-file-path-v1 (-> Url Path-String)) -(define (url->cache-file-path-v1 uri) +(define (url->cache-file-path-v1 url) (define (hash-sha1 str) : (-> String String) (define in (open-input-string str)) (define digest (sha1 in)) (close-input-port in) digest) - (build-path cache-object-dir (hash-sha1 (url->string uri)))) + (build-path cache-object-dir (hash-sha1 (url->string url)))) (: url->cache-file-path-v2 (-> Url Path-String)) -(define (url->cache-file-path-v2 uri) - (build-path cache-object-dir (uri-encode (url->string uri)))) +(define (url->cache-file-path-v2 url) + (build-path cache-object-dir (uri-encode (url->string url)))) (define url->cache-object-path url->cache-file-path-v2) -(define (url->cache-etag-path uri) - (build-path cache-dir "etags" (uri-encode (url->string uri)))) +(define (url->cache-etag-path url) + (build-path cache-dir "etags" (uri-encode (url->string url)))) -(define (url->cache-lmod-path uri) - (build-path cache-dir "lmods" (uri-encode (url->string uri)))) +(define (url->cache-lmod-path url) + (build-path cache-dir "lmods" (uri-encode (url->string url)))) -(: uri-read-cached (-> Url (Option String))) -(define (uri-read-cached uri) - (define path-v1 (url->cache-file-path-v1 uri)) - (define path-v2 (url->cache-file-path-v2 uri)) +(: url-read-cached (-> Url (Option String))) +(define (url-read-cached url) + (define path-v1 (url->cache-file-path-v1 url)) + (define path-v2 (url->cache-file-path-v2 url)) (when (file-exists? path-v1) (rename-file-or-directory path-v1 path-v2 #t)) (if (file-exists? path-v2) (file->string path-v2) (begin - (log-debug "Cache file not found for URI: ~a" (url->string uri)) + (log-debug "Cache file not found for URL: ~a" (url->string url)) #f))) (: str->url (-> String (Option String))) @@ -479,7 +477,7 @@ comment) (match (str->url url) [#f - (log-error "Invalid URI in peer string: ~v" str) + (log-error "Invalid URL in peer string: ~v" str) #f] [url (Peer nick url (url->string url) comment)])] @@ -588,15 +586,15 @@ [(list val) val] [_ #f])) -(: uri-download-from-port +(: url-download-http-from-port (-> Url (Listof (U Bytes String)) Input-Port (U 'skipped-cached 'downloaded-new))) ; TODO 'ok|'error ? -(define (uri-download-from-port u headers body-input) +(define (url-download-http-from-port u headers body-input) ; TODO Update message db from here? or where? ; - 1st try can just be an in-memory set that gets written-to ; and read-from disk as a whole. (define u-str (url->string u)) - (log-debug "uri-download-from-port ~v into ~v" u-str cached-object-path) + (log-debug "url-download-http-from-port ~v into ~v" u-str cached-object-path) (define cached-object-path (url->cache-object-path u)) (define cached-etag-path (url->cache-etag-path u)) (define cached-lmod-path (url->cache-lmod-path u)) @@ -638,11 +636,8 @@ 'downloaded-new) 'skipped-cached)) -(: uri-download - (-> Positive-Float Url - (Result (U 'skipped-cached 'downloaded-new) - Any))) ; TODO Maybe more-precise error type? -(define (uri-download timeout u) +(: url-download-http (-> Positive-Float Url Download-Result)) +(define (url-download-http timeout u) (define u-str (url->string u)) (define timeout-chan (make-channel)) (define result-chan (make-channel)) @@ -675,7 +670,7 @@ ; TODO Should a redirect update a peer URL? (match status [200 - `(ok . ,(uri-download-from-port u headers body-input))] + `(ok . ,(url-download-http-from-port u headers body-input))] [_ `(error . (http-not-ok . ,status))])]) (close-input-port body-input) @@ -686,6 +681,15 @@ (kill-thread timeout-thread) result) +(: url-download (-> Positive-Float Url Download-Result)) +(define (url-download timeout u) + (match (url-scheme u) + ; TODO Support Gopher. + [(or "http" "https") + (url-download-http timeout u)] + [scheme + `(error . (unsupported-url-scheme . ,scheme))])) + (: timeline-print (-> Out-Format (Listof Msg) Void)) (define (timeline-print out-format timeline) (match timeline @@ -702,9 +706,9 @@ (: peer->msgs (-> Peer (Listof Msg))) (define (peer->msgs peer) - (match-define (Peer nick uri uri-str _) peer) - (log-debug "Reading peer nick:~v uri:~v" nick uri-str) - (define msgs-data (uri-read-cached uri)) + (match-define (Peer nick url url-str _) peer) + (log-debug "Reading peer nick:~v url:~v" nick url-str) + (define msgs-data (url-read-cached url)) ; TODO Expire cache (if msgs-data (str->msgs peer msgs-data) @@ -715,10 +719,10 @@ (Result (U 'skipped-cached 'downloaded-new) Any))) (define (peer-download timeout peer) - (match-define (Peer nick uri u _) peer) + (match-define (Peer nick url u _) peer) (log-info "Download BEGIN URL:~a" u) (define-values (results _tm-cpu-ms tm-real-ms _tm-gc-ms) - (time-apply uri-download (list timeout uri))) + (time-apply url-download (list timeout url))) (define result (car results)) (log-info "Download END in ~a seconds, URL:~a, result:~s" (/ tm-real-ms 1000.0) @@ -801,7 +805,7 @@ (set-map denied-hosts (λ (h) (pregexp (string-append "\\." h "$"))))) (filter (λ (p) - (define host (url-host (Peer-uri p))) + (define host (url-host (Peer-url p))) (not (or (set-member? denied-hosts host) (ormap (λ (d) (regexp-match? d host)) denied-domain-patterns)))) peers)) @@ -830,7 +834,7 @@ (current-logger logger) log-writer)) -(: msgs->nick-hist (-> (Listof Msg) Nick-Hist)) +(: msgs->nick-hist (-> (Listof Msg) Url-Nick-Hist)) (define (msgs->nick-hist msgs) (foldl (λ (msg url->nick->hist) @@ -855,13 +859,33 @@ (hash) msgs)) -(: update-nicks-history-files (-> Nick-Hist Void)) -(define (update-nicks-history-files nick-hist) +(: url-nick-hist->file (-> Url-Nick-Hist Path-String Void)) +(define (url-nick-hist->file unh filepath) + (define out (open-output-file filepath #:exists 'replace)) + (for-each + (match-lambda + [(cons url nick->hist) + (displayln (url->string url) out) + (for-each (match-lambda + [(cons nick (Hist freq last)) + (displayln (format " ~a ~a ~a" nick freq last) out)]) + (sort (hash->list nick->hist) + (match-lambda** + [((cons _ (Hist a _)) (cons _ (Hist b _))) + (> a b)])))]) + (sort + (hash->list unh) + (λ (a b) (stringdir (-> Url-Nick-Hist Path-String Void)) +(define (url-nick-hist->dir unh dirpath) (hash-for-each - nick-hist + unh (λ (url nick->hist) (define filename (string-append (uri-encode (url->string url)) ".txt")) - (define filepath (build-path tt-home-dir "nicks" "seen" filename)) + (define filepath (build-path dirpath filename)) (make-parent-directory* filepath) (display-lines-to-file (map (match-lambda @@ -874,8 +898,14 @@ filepath #:exists 'replace)))) -(: nick-hist-most-by (-> Nick-Hist Url (-> Hist Nonnegative-Integer) (Option String))) -(define (nick-hist-most-by url->nick->hist url by) +(: update-nicks-history-files (-> Url-Nick-Hist Void)) +(define (update-nicks-history-files unh) + (define nicks-dir (build-path tt-home-dir "nicks")) + (url-nick-hist->file unh (build-path nicks-dir "seen.txt")) + (url-nick-hist->dir unh (build-path nicks-dir "seen"))) + +(: url-nick-hist-most-by (-> Url-Nick-Hist Url (-> Hist Nonnegative-Integer) (Option String))) +(define (url-nick-hist-most-by url->nick->hist url by) (match (hash-ref url->nick->hist url #f) [#f #f] [nick->hist @@ -885,19 +915,19 @@ ['() #f] [(cons (cons nick _) _) nick])])) -(: nick-hist-latest (-> Nick-Hist Url (Option String))) -(define (nick-hist-latest nick-hist url) - (nick-hist-most-by nick-hist url Hist-last)) +(: url-nick-hist-latest (-> Url-Nick-Hist Url (Option String))) +(define (url-nick-hist-latest unh url) + (url-nick-hist-most-by unh url Hist-last)) -(: nick-hist-common (-> Nick-Hist Url (Option String))) -(define (nick-hist-common nick-hist url) - (nick-hist-most-by nick-hist url Hist-freq)) +(: url-nick-hist-common (-> Url-Nick-Hist Url (Option String))) +(define (url-nick-hist-common unh url) + (url-nick-hist-most-by unh url Hist-freq)) -(: peers-update-nick-to-common (-> Nick-Hist (Listof Peer) (Listof Peer))) -(define (peers-update-nick-to-common nick-hist peers) +(: peers-update-nick-to-common (-> Url-Nick-Hist (Listof Peer) (Listof Peer))) +(define (peers-update-nick-to-common unh peers) (map (λ (p) - (match (nick-hist-common nick-hist (Peer-uri p)) + (match (url-nick-hist-common unh (Peer-url p)) [#f p] [n (struct-copy Peer p [nick n])])) peers)) @@ -928,8 +958,8 @@ (check-equal? (hash-ref (hash-ref hist url) nick1) (Hist 1 ts-2)) (check-equal? (hash-ref (hash-ref hist url) nick2) (Hist 2 ts-1)) (check-equal? (hash-ref (hash-ref hist url) nick3) (Hist 3 ts-1)) - (check-equal? (nick-hist-common hist url) nick3) - (check-equal? (nick-hist-latest hist url) nick1))) + (check-equal? (url-nick-hist-common hist url) nick3) + (check-equal? (url-nick-hist-latest hist url) nick1))) (: crawl (-> Void)) (define (crawl) @@ -946,7 +976,7 @@ (peers-cached)] [cached-timeline (peers->timeline peers-cached)] - [nick-hist + [url-nick-hist (msgs->nick-hist cached-timeline)] [peers-mentioned-curr (peers-mentioned cached-timeline)] @@ -959,7 +989,7 @@ peers-mentioned-curr)] [peers-all (peers-update-nick-to-common - nick-hist + url-nick-hist (peers-merge peers-mentioned peers-all-prev peers-cached))] @@ -979,7 +1009,7 @@ (match-lambda [(Peer n _ u c) (list n u c)]) peers-discovered))) - (update-nicks-history-files nick-hist) + (update-nicks-history-files url-nick-hist) (peers->file peers-cached peers-cached-file) (peers->file peers-mentioned @@ -1024,17 +1054,22 @@ (define (dispatch command) (match command [(or "d" "download") - (let ([num-workers 20] ; 20 was fastest out of the tried: 1, 5, 10, 20, 25, 30. - [timeout 10.0]) + ; 20 was fastest out of the tried: 1, 5, 10, 20, 25, 30. + (let ([num-workers : Positive-Integer 20] + [timeout : Positive-Flonum 10.0]) (command-line #:program "tt download" #:once-each [("-j" "--jobs") - njobs "Number of concurrent jobs." - (set! num-workers (string->number njobs))] + positive-integer "Number of concurrent jobs." + (set! num-workers + (assert (string->number positive-integer) + (conjoin exact-positive-integer?)))] [("-t" "--timeout") - seconds "Timeout seconds per request." - (set! timeout (string->number seconds))] + positive-float "Timeout seconds per request." + (set! timeout + (assert (string->number positive-float) + (conjoin positive? flonum?)))] #:args file-paths (download file-paths num-workers timeout)))] [(or "u" "upload")