X-Git-Url: https://git.xandkar.net/?a=blobdiff_plain;f=tt.rkt;h=e93c26f4a20e67426ec608c863204e638dead67f;hb=HEAD;hp=335fa6c83b1d79b2c3b15d124e590591b617a247;hpb=0cb1ae9c9a4d1052ee1b7a5b540e2f261d53153c;p=tt.git diff --git a/tt.rkt b/tt.rkt index 335fa6c..e93c26f 100644 --- a/tt.rkt +++ b/tt.rkt @@ -27,6 +27,34 @@ (∀ (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 Url-Nick-Hist + (Immutable-HashTable Url (Immutable-HashTable (Option String) Hist))) + +(struct User + ([url : Url] + [nick : (Option String)])) + +(struct User-Agent + ([user : User] + [prog : Prog])) + +(struct Prog + ([name : String] + [version : String])) + (struct Msg ([ts-epoch : Integer] [ts-orig : String] @@ -36,25 +64,87 @@ (struct Peer ([nick : (Option String)] - [uri : Url] - [uri-str : String] + [url : Url] + [url-str : String] [comment : (Option String)]) #:transparent) -(struct Resp - ([status-line : String] - [headers : (Listof Bytes)] - [body-input : Input-Port]) - #:transparent) +(: prog Prog) +(define prog + (Prog "tt" (info:#%info-lookup 'version))) + +(: user-default User) +(define user-default + (User (string->url "https://github.com/xandkar/tt") #f)) + +(: user->str (-> User String)) +(define (user->str user) + (match-define (User u0 n) user) + (define u (url->string u0)) + (if n + (format "+~a; @~a" u n) + (format "+~a" u ))) + +(: user-agent->str (-> User-Agent String)) +(define (user-agent->str ua) + (match-define (User-Agent u p) ua) + (format "~a/~a (~a)" (Prog-name p) (Prog-version p) (user->str u))) + +(: user->user-agent User) +(define (user->user-agent user) + (User-Agent user prog)) + +(: user-agent-str String) +(define user-agent-str + (user-agent->str (user->user-agent user-default))) + +(: set-user-agent-str (-> Path-String Void)) +(define (set-user-agent-str filename) + (set! user-agent-str (user-agent->str (user->user-agent (file->user filename)))) + (log-info "User-Agent string is now set to: ~v" user-agent-str)) + +(: file->user (-> Path-String User)) +(define (file->user filename) + (if (file-exists? filename) + (match (file->peers filename) + [(list p) + (log-info + "User-Agent. Found one peer in file: ~v. Using the found peer: ~a" + filename + (peer->str p)) + (peer->user p)] + [(list* p _) + (log-warning + "User-Agent. Multiple peers in file: ~v. Picking arbitrary: ~a" + filename + (peer->str p)) + (peer->user p)] + ['() + (log-warning + "User-Agent. No peers found in file: ~v. Using the default user: ~a" + filename + user-default) + user-default]) + (begin + (log-warning + "User-Agent. File doesn't exist: ~v. Using the default user: ~a" + filename + user-default) + user-default))) + +(: peer->user (-> Peer User)) +(define (peer->user p) + (match-define (Peer n u _ _) p) + (User u n)) (: 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? @@ -74,9 +164,41 @@ ; ; TODO Investigate why and make a minimal reproducible test case. +(: peers-merge (-> (Listof Peer) * (Listof Peer))) +(define (peers-merge . peer-sets) + (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) (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 _)) (stringurl u1) u1 #f)] + [p2 (Peer "a" (string->url u1) u1 #f)] + [p3 (Peer "b" (string->url u2) u2 #f)] + [s1 (list p1)] + [s2 (list p2 p3)]) + (check-equal? (list p3 p2) (peers-merge s1 s2)) + (check-equal? (list p3 p2) (peers-merge s2 s1)))) + (: tt-home-dir Path-String) (define tt-home-dir (build-path (expand-user-path "~") ".tt")) +(: pub-peers-dir Path-String) +(define pub-peers-dir (build-path tt-home-dir "peers")) + (: concurrent-filter-map (∀ (α β) (-> Natural (-> α β) (Listof α) (Listof β)))) (define (concurrent-filter-map num-workers f xs) ; TODO preserve order of elements OR communicate that reorder is expected @@ -129,11 +251,11 @@ (λ (out-format color-i msg) (let ([color (vector-ref colors (modulo color-i n))] [nick (Peer-nick (Msg-from msg))] - [uri (Peer-uri-str (Msg-from msg))] + [url (Peer-url-str (Msg-from msg))] [text (Msg-text msg)]) (match out-format ['single-line - (let ([nick (if nick nick uri)]) + (let ([nick (if nick nick url)]) (printf "~a \033[1;37m<~a>\033[0m \033[0;~am~a\033[0m~n" (parameterize ([date-display-format 'iso-8601]) @@ -146,7 +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 @@ -202,8 +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 @@ -223,8 +345,8 @@ (string-append d h m))] [tzs (list* "" "Z" tzs)]) (for* ([n '("fake-nick")] - [u '("http://fake-uri")] - [p (list (Peer n (string->url u) u ""))] + [u '("http://fake-url")] + [p (list (Peer n (string->url u) u #f))] [s '("" ":10")] [f '("" ".1337")] [z tzs] @@ -244,8 +366,8 @@ [tab " "] [text "Lorem ipsum"] [nick "foo"] - [uri "http://bar/"] - [peer (Peer nick (string->url uri) uri "")] + [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? @@ -261,13 +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) @@ -280,6 +402,8 @@ (module+ test (check-equal? (str->lines "abc\ndef\n\nghi") '("abc" "def" "ghi"))) +; TODO Should return 2 things: 1) msgs; 2) metadata parsed from comments +; TODO Update peer nick based on metadata? (: str->msgs (-> Peer String (Listof Msg))) (define (str->msgs peer str) (filter-map (λ (line) (str->msg peer line)) @@ -291,37 +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))) @@ -353,7 +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)])] @@ -394,15 +518,16 @@ (define (filter-comments lines) (filter-not (λ (line) (string-prefix? line "#")) lines)) -(: str->peers (-> String (Setof Peer))) +(: str->peers (-> String (Listof Peer))) (define (str->peers str) - (make-immutable-peers (filter-map str->peer (filter-comments (str->lines str))))) + (filter-map str->peer (filter-comments (str->lines str)))) -(: peers->file (-> (Setof Peers) Path-String Void)) +(: peers->file (-> (Listof Peers) Path-String Void)) (define (peers->file peers path) + (make-parent-directory* path) (display-lines-to-file (map peer->str - (sort (set->list peers) + (sort peers (match-lambda** [((Peer n1 _ _ _) (Peer n2 _ _ _)) (stringpeers (-> Path-String (Setof Peer))) +(: file->peers (-> Path-String (Listof Peer))) (define (file->peers file-path) (if (file-exists? file-path) (str->peers (file->string file-path)) (begin (log-warning "File does not exist: ~v" (path->string file-path)) - (make-immutable-peers)))) + '()))) (define re-rfc2822 #px"^(Mon|Tue|Wed|Thu|Fri|Sat|Sun), ([0-9]{2}) (Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec) ([0-9]{4}) ([0-2][0-9]):([0-6][0-9]):([0-6][0-9]) GMT") @@ -455,33 +580,21 @@ [_ #f])) -(: user-agent String) -(define user-agent - (let* - ([prog-name "tt"] - [prog-version (info:#%info-lookup 'version)] - [prog-uri "https://github.com/xandkar/tt"] - [user-peer-file (build-path tt-home-dir "me")] - [user - (if (file-exists? user-peer-file) - (match (set-first (file->peers user-peer-file)) - [(Peer #f _ u _) (format "+~a" u )] - [(Peer n _ u _) (format "+~a; @~a" u n)]) - (format "+~a" prog-uri))]) - (format "~a/~a (~a)" prog-name prog-version user))) - (: header-get (-> (Listof Bytes) Bytes (Option Bytes))) (define (header-get headers name) (match (filter-map (curry extract-field name) headers) [(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)) @@ -523,11 +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)) @@ -539,19 +649,6 @@ (channel-put timeout-chan '(error . timeout))))) (define result-thread (thread (λ () - ; XXX We timeout getting a response, but body download could - ; also take a long time and we might want to time that out as - ; well, but then we may end-up with partially downloaded - ; objects. But that could happen anyway if the server drops the - ; connection for whatever reason. - ; - ; Maybe that is OK once we start treating the - ; downloaded object as an addition to the stored set of - ; messages, rather than the final set of messages. - - ; TODO message db - ; - 1st try can just be an in-memory set that gets written-to - ; and read-from disk as a whole. (define result (with-handlers ; TODO Maybe name each known errno? (exn:fail:network:errno-errno e) @@ -562,32 +659,36 @@ (define-values (status-line headers body-input) (http-sendrecv/url u - #:headers (list (format "User-Agent: ~a" user-agent)))) - `(ok . ,(Resp status-line headers body-input)))) + #:headers (list (format "User-Agent: ~a" user-agent-str)))) + (log-debug "headers: ~v" headers) + (log-debug "status-line: ~v" status-line) + (define status + (string->number (second (string-split (bytes->string/utf-8 status-line))))) + (log-debug "status: ~v" status) + (let ([result + ; TODO Handle redirects. + ; TODO Should a redirect update a peer URL? + (match status + [200 + `(ok . ,(url-download-http-from-port u headers body-input))] + [_ + `(error . (http-not-ok . ,status))])]) + (close-input-port body-input) + result))) (channel-put result-chan result)))) - (define result - (sync timeout-chan - result-chan)) + (define result (sync timeout-chan result-chan)) (kill-thread result-thread) (kill-thread timeout-thread) - (match result - [(cons 'error _) - result] - [(cons 'ok (Resp status-line headers body-input)) - (log-debug "headers: ~v" headers) - (log-debug "status-line: ~v" status-line) - (define status - (string->number (second (string-split (bytes->string/utf-8 status-line))))) - (log-debug "status: ~v" status) - ; TODO Handle redirects. Should be within same timeout as req and body. - (let ([result - (match status - [200 - `(ok . ,(uri-download-from-port u headers body-input))] - [_ - `(error . (http . ,status))])]) - (close-input-port body-input) - result)])) + 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) @@ -605,9 +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) @@ -618,10 +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) @@ -629,12 +730,12 @@ result) result) -(: timeline-download (-> Integer Positive-Float (Setof Peer) Void)) +(: timeline-download (-> Integer Positive-Float (Listof Peer) Void)) (define (timeline-download num-workers timeout peers) (define results (concurrent-filter-map num-workers (λ (p) (cons p (peer-download timeout p))) - (set->list peers))) + peers)) (define peers-ok (filter-map (match-lambda [(cons p (cons 'ok _)) p] @@ -647,12 +748,12 @@ [(cons p (cons 'error e)) (struct-copy Peer p [comment (format "~s" e)])]) results)) - (peers->file peers-ok (build-path tt-home-dir "peers-last-downloaded-ok")) - (peers->file peers-err (build-path tt-home-dir "peers-last-downloaded-err"))) + (peers->file peers-ok (build-path tt-home-dir "peers-last-downloaded-ok.txt")) + (peers->file peers-err (build-path tt-home-dir "peers-last-downloaded-err.txt"))) (: peers->timeline (-> (Listof Peer) (Listof Msg))) (define (peers->timeline peers) - (append* (filter-map peer->msgs (set->list peers)))) + (append* (filter-map peer->msgs peers))) (: timeline-sort (-> (Listof Msg) timeline-order (Listof Msgs))) (define (timeline-sort msgs order) @@ -662,11 +763,11 @@ (sort msgs (λ (a b) (cmp (Msg-ts-epoch a) (Msg-ts-epoch b))))) -(: paths->peers (-> (Listof String) (Setof Peer))) +(: paths->peers (-> (Listof String) (Listof Peer))) (define (paths->peers paths) (let* ([paths (match paths ['() - (let ([peer-refs-file (build-path tt-home-dir "peers")]) + (let ([peer-refs-file (build-path tt-home-dir "following.txt")]) (log-debug "No peer ref file paths provided, defaulting to ~v" (path->string peer-refs-file)) @@ -674,39 +775,40 @@ [paths (log-debug "Peer ref file paths provided: ~v" paths) (map string->path paths)])] - [peers (apply set-union (map file->peers paths))]) - (log-info "Read-in ~a peers." (set-count peers)) + [peers (apply peers-merge (map file->peers paths))]) + (log-info "Read-in ~a peers." (length peers)) peers)) -(: mentioned-peers-in-cache (-> (Setof Peer))) -(define (mentioned-peers-in-cache) - ; TODO Expire cache - (define msgs - (append* (map (λ (filename) - (define path (build-path cache-object-dir filename)) - (define size (/ (file-size path) 1000000.0)) - (log-debug "BEGIN parsing ~a MB from file: ~v" - size - (path->string path)) - (define t0 (current-inexact-milliseconds)) - (define m (filter-map - (λ (line) - (define url-str (uri-decode (path->string filename))) - (define url (string->url url-str)) - (define from (Peer #f url url-str "")) - (str->msg from line)) - (filter-comments - (file->lines path)))) - (define t1 (current-inexact-milliseconds)) - (log-debug "END parsing ~a MB in ~a seconds from file: ~v." - size - (* 0.001 (- t1 t0)) - (path->string path)) - (when (empty? m) - (log-debug "No messages found in ~a" (path->string path))) - m) - (directory-list cache-object-dir)))) - (make-immutable-peers (append* (map Msg-mentions msgs)))) +(: cache-filename->peer (-> Path-String (Option Peer))) +(define (cache-filename->peer filename) + (define nick #f) ; TODO Look it up in the nick-db when it exists. + (define url-str (uri-decode (path->string filename))) ; TODO Can these crash? + (match (str->url url-str) + [#f #f] + [url (Peer nick url url-str #f)])) + +(: peers-cached (-> (Listof Peer))) +(define (peers-cached) + ; TODO Expire cache? + (filter-map cache-filename->peer (directory-list cache-object-dir))) + +(: peers-mentioned (-> (Listof Msg) (Listof Peer))) +(define (peers-mentioned msgs) + (append* (map Msg-mentions msgs))) + +(: peers-filter-denied-domains (-> (Listof Peer) (Listof Peer))) +(define (peers-filter-denied-domains peers) + (define deny-file (build-path tt-home-dir "domains-deny.txt")) + (define denied-hosts + (list->set (map string-trim (filter-comments (file->lines deny-file))))) + (define denied-domain-patterns + (set-map denied-hosts (λ (h) (pregexp (string-append "\\." h "$"))))) + (filter + (λ (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)) (: log-writer-stop (-> Thread Void)) (define (log-writer-stop log-writer) @@ -732,41 +834,184 @@ (current-logger logger) log-writer)) +(: msgs->nick-hist (-> (Listof Msg) Url-Nick-Hist)) +(define (msgs->nick-hist msgs) + (foldl + (λ (msg url->nick->hist) + (match-define (Msg curr _ from _ mentions) msg) + (foldl + (λ (peer url->nick->hist) + (match-define (Peer nick url _ _) peer) + (if nick + (hash-update url->nick->hist + url + (λ (nick->hist) + (hash-update nick->hist + nick + (match-lambda + [(Hist freq prev) + (Hist (+ 1 freq) (max prev curr))]) + (Hist 0 0))) + (hash)) + url->nick->hist)) + url->nick->hist + (cons from mentions))) + (hash) + msgs)) + +(: 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 + unh + (λ (url nick->hist) + (define filename (string-append (uri-encode (url->string url)) ".txt")) + (define filepath (build-path dirpath filename)) + (make-parent-directory* filepath) + (display-lines-to-file + (map (match-lambda + [(cons nick (Hist freq last)) + (format "~a ~a ~a" nick freq last)]) + (sort (hash->list nick->hist) + (match-lambda** + [((cons _ (Hist a _)) (cons _ (Hist b _))) + (> a b)]))) + filepath + #:exists 'replace)))) + +(: 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 + (match (sort (hash->list nick->hist) + (λ (a b) (> (by (cdr a)) + (by (cdr b))))) + ['() #f] + [(cons (cons nick _) _) nick])])) + +(: 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)) + +(: 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 (-> Url-Nick-Hist (Listof Peer) (Listof Peer))) +(define (peers-update-nick-to-common unh peers) + (map + (λ (p) + (match (url-nick-hist-common unh (Peer-url p)) + [#f p] + [n (struct-copy Peer p [nick n])])) + peers)) + +(module+ test + (let* ([url-str "http://foo"] + [url (string->url url-str)] + [nick1 "a"] + [nick2 "b"] + [nick3 "c"] + [ts-str-1 "2021-11-29T23:29:08-0500"] + [ts-str-2 "2021-11-29T23:30:00-0500"] + [ts-1 (rfc3339->epoch ts-str-1)] + [ts-2 (rfc3339->epoch ts-str-2)] + [msgs + (map (match-lambda + [(cons ts-str nick) + (str->msg (str->peer "test http://test") + (string-append ts-str " Hi @<" nick " " url-str ">"))]) + (list (cons ts-str-2 nick1) + (cons ts-str-1 nick2) + (cons ts-str-1 nick2) + (cons ts-str-1 nick3) + (cons ts-str-1 nick3) + (cons ts-str-1 nick3)))] + [hist + (msgs->nick-hist msgs)]) + (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? (url-nick-hist-common hist url) nick3) + (check-equal? (url-nick-hist-latest hist url) nick1))) + (: crawl (-> Void)) (define (crawl) + ; TODO Test the non-io parts of crawling (let* ([peers-all-file - (build-path tt-home-dir "peers-all")] + (build-path pub-peers-dir "all.txt")] [peers-mentioned-file - (build-path tt-home-dir "peers-mentioned")] + (build-path pub-peers-dir "mentioned.txt")] [peers-parsed-file - (build-path tt-home-dir "peers-parsed")] + (build-path pub-peers-dir "downloaded-and-parsed.txt")] + [peers-cached-file + (build-path pub-peers-dir "downloaded.txt")] + [peers-cached + (peers-cached)] + [cached-timeline + (peers->timeline peers-cached)] + [url-nick-hist + (msgs->nick-hist cached-timeline)] [peers-mentioned-curr - (mentioned-peers-in-cache)] + (peers-mentioned cached-timeline)] [peers-mentioned-prev (file->peers peers-mentioned-file)] - [peers-mentioned - (set-union peers-mentioned-prev - peers-mentioned-curr)] [peers-all-prev (file->peers peers-all-file)] + [peers-mentioned + (peers-merge peers-mentioned-prev + peers-mentioned-curr)] [peers-all - (set-union peers-mentioned - peers-all-prev)] + (peers-update-nick-to-common + url-nick-hist + (peers-merge peers-mentioned + peers-all-prev + peers-cached))] [peers-discovered - (set-subtract peers-all - peers-all-prev)] + (set->list (set-subtract (make-immutable-peers peers-all) + (make-immutable-peers peers-all-prev)))] [peers-parsed - (for/set ([p peers-all] #:when (> (length (peer->msgs p)) 0)) p)]) + (filter (λ (p) (> (length (peer->msgs p)) 0)) peers-all)]) ; TODO Deeper de-duping - (log-info "Known peers mentioned: ~a" (set-count peers-mentioned)) - (log-info "Known peers parsed ~a" (set-count peers-parsed)) - (log-info "Known peers total: ~a" (set-count peers-all)) + (log-info "Known peers cached ~a" (length peers-cached)) + (log-info "Known peers mentioned: ~a" (length peers-mentioned)) + (log-info "Known peers parsed ~a" (length peers-parsed)) + (log-info "Known peers total: ~a" (length peers-all)) (log-info "Discovered ~a new peers:~n~a" - (set-count peers-discovered) + (length peers-discovered) (pretty-format (map (match-lambda [(Peer n _ u c) (list n u c)]) - (set->list peers-discovered)))) + peers-discovered))) + (update-nicks-history-files url-nick-hist) + (peers->file peers-cached + peers-cached-file) (peers->file peers-mentioned peers-mentioned-file) (peers->file peers-parsed @@ -795,30 +1040,36 @@ (: download (-> (Listof String) Positive-Integer Positive-Float Void)) (define (download file-paths num-workers timeout) - (let ([peers (paths->peers file-paths)]) + (let* ([peers-given (paths->peers file-paths)] + [peers-kept (peers-filter-denied-domains peers-given)] + [peers-denied (set-subtract peers-given peers-kept)]) + (log-info "Denied ~a peers" (length peers-denied)) (define-values (_res _cpu real-ms _gc) - (time-apply timeline-download (list num-workers timeout peers))) + (time-apply timeline-download (list num-workers timeout peers-kept))) (log-info "Downloaded timelines from ~a peers in ~a seconds." - (set-count peers) + (length peers-kept) (/ real-ms 1000.0)))) (: dispatch (-> String Void)) (define (dispatch command) (match command [(or "d" "download") - ; Initially, 15 was fastest out of the tried: 1, 5, 10, 20. Then I - ; started noticing significant slowdowns. Reducing to 5 seems to help. - (let ([num-workers 5] - [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") @@ -879,6 +1130,7 @@ #:args (command . args) (define log-writer (log-writer-start log-level)) (current-command-line-arguments (list->vector args)) + (set-user-agent-str (build-path tt-home-dir "user.txt")) ; TODO dispatch should return status with which we should exit after cleanups (dispatch command) (log-writer-stop log-writer))))