Implement user-enabled, shorter short format (no URI)
[tt.git] / tt.rkt
1 ; TODO optional text wrap
2 ; TODO write
3 ; TODO caching (use cache by default, unless explicitly asked for update)
4 ; - [x] value --> cache
5 ; - [x] value <-- cache
6 ; requires: commands
7 ; TODO timeline limits
8 ; TODO feed set operations (perhaps better done externally?)
9 ; TODO timeline as a result of a query (feed set op + filter expressions)
10 ; TODO named timelines
11 ; TODO config files
12 ; TODO parse "following" from feed
13 ; - following = <nick> <uri>
14 ; TODO parse mentions:
15 ; - @<source.nick source.url> | @<source.url>
16 ; TODO highlight mentions
17 ; TODO filter on mentions
18 ; TODO highlight hashtags
19 ; TODO filter on hashtags
20 ; TODO hashtags as channels? initial hashtag special?
21 ; TODO query language
22 ; TODO console logger colors by level ('error)
23 ; TODO file logger ('debug)
24 ; TODO commands:
25 ; - r | read
26 ; - see timeline ops above
27 ; - w | write
28 ; - arg or stdin
29 ; - nick expand to URI
30 ; - q | query
31 ; - see timeline ops above
32 ; - see hashtag and channels above
33 ; - d | download
34 ; - u | upload
35 ; - calls user-configured command to upload user's own feed file to their server
36 ;
37 ; Looks like a better CLI parser than "racket/cmdline":
38 ; https://docs.racket-lang.org/natural-cli/
39
40 #lang racket
41
42 (require openssl/sha1)
43 (require racket/date)
44
45 (require http-client)
46 (require rfc3339-old)
47
48 (module+ test
49 (require rackunit))
50
51 (struct msg (ts_epoch ts_rfc3339 nick uri text))
52 (struct feed (nick uri))
53
54 (define (concurrent-filter-map num_workers f xs)
55 ; TODO preserve order of elements OR communicate that reorder is expected
56 ; TODO switch from mailboxes to channels
57 (define (make-worker id f)
58 (define parent (current-thread))
59 (λ ()
60 (define self (current-thread))
61 (define (work)
62 (thread-send parent (cons 'next self))
63 (match (thread-receive)
64 ['done (thread-send parent (cons 'exit id))]
65 [(cons 'unit x) (begin
66 (define y (f x))
67 (when y (thread-send parent (cons 'result y)))
68 (work))]))
69 (work)))
70 (define (dispatch ws xs ys)
71 (if (empty? ws)
72 ys
73 (match (thread-receive)
74 [(cons 'exit w) (dispatch (remove w ws =) xs ys)]
75 [(cons 'result y) (dispatch ws xs (cons y ys))]
76 [(cons 'next thd) (match xs
77 ['() (begin
78 (thread-send thd 'done)
79 (dispatch ws xs ys))]
80 [(cons x xs) (begin
81 (thread-send thd (cons 'unit x))
82 (dispatch ws xs ys))])])))
83 (define workers (range num_workers))
84 (define threads (map (λ (id) (thread (make-worker id f))) workers))
85 (define results (dispatch workers xs '()))
86 (for-each thread-wait threads)
87 results)
88
89 (module+ test
90 (let* ([f (λ (x) (if (even? x) x #f))]
91 [xs (range 11)]
92 [actual (sort (concurrent-filter-map 10 f xs) <)]
93 [expected (sort ( filter-map f xs) <)])
94 (check-equal? actual expected "concurrent-filter-map")))
95
96 (define msg-print
97 (let* ([colors (vector 36 33)]
98 [n (vector-length colors)])
99 (λ (out-format color-i msg)
100 (let ([color (vector-ref colors (modulo color-i n))]
101 [nick (msg-nick msg)]
102 [uri (msg-uri msg)]
103 [text (msg-text msg)])
104 (match out-format
105 ['single-line
106 (printf "~a \033[1;37m<~a>\033[0m \033[0;~am~a\033[0m~n"
107 (parameterize ([date-display-format 'iso-8601])
108 (date->string (seconds->date [msg-ts_epoch msg]) #t))
109 nick color text)]
110 ['multi-line
111 (printf "~a~n\033[1;37m<~a ~a>\033[0m~n\033[0;~am~a\033[0m~n~n"
112 (parameterize ([date-display-format 'rfc2822])
113 (date->string (seconds->date [msg-ts_epoch msg]) #t))
114 nick uri color text)])))))
115
116 (define re-msg-begin
117 ; TODO Zulu offset. Maybe in several formats. Which ones?
118 (pregexp "^[0-9]{4}-[0-9]{2}-[0-9]{2}T[0-9]{2}:[0-9]{2}:[0-9]{2}"))
119
120 (define (str->msg nick uri str)
121 (if (not (regexp-match? re-msg-begin str))
122 (begin
123 (log-debug "Non-msg line from nick:~a, line:~a" nick str)
124 #f)
125 (let ([toks (string-split str (regexp "\t+"))])
126 (if (not (= 2 (length toks)))
127 (begin
128 (log-warning "Invalid msg line from nick:~a, msg:~a" nick str)
129 #f)
130 (let*
131 ([ts_rfc3339 (first toks)]
132 [text (second toks)]
133 [t (string->rfc3339-record ts_rfc3339)]
134 ; TODO handle tz offset
135 [ts_epoch (find-seconds [rfc3339-record:second t]
136 [rfc3339-record:minute t]
137 [rfc3339-record:hour t]
138 [rfc3339-record:mday t]
139 [rfc3339-record:month t]
140 [rfc3339-record:year t])])
141 (msg ts_epoch ts_rfc3339 nick uri text))))))
142
143 (module+ test
144 (let* ([ts "2020-11-18T22:22:09-0500"]
145 [tab " "]
146 [text "Lorem ipsum"]
147 [nick "foo"]
148 [uri "bar"]
149 [actual (str->msg nick uri (string-append ts tab text))]
150 [expected (msg 1605756129 ts nick uri text)])
151 ; FIXME re-enable after handling tz offset
152 ;(check-equal?
153 ; (msg-ts_epoch actual)
154 ; (msg-ts_epoch expected)
155 ; "str->msg ts_epoch")
156 (check-equal?
157 (msg-ts_rfc3339 actual)
158 (msg-ts_rfc3339 expected)
159 "str->msg ts_rfc3339")
160 (check-equal?
161 (msg-nick actual)
162 (msg-nick expected)
163 "str->msg nick")
164 (check-equal?
165 (msg-uri actual)
166 (msg-uri expected)
167 "str->msg uri")
168 (check-equal?
169 (msg-text actual)
170 (msg-text expected)
171 "str->msg text")))
172
173 (define (str->lines str)
174 (string-split str (regexp "[\r\n]+")))
175
176 (module+ test
177 (check-equal? (str->lines "abc\ndef\n\nghi") '("abc" "def" "ghi")))
178
179 (define (str->msgs nick uri str)
180 (filter-map (λ (line) (str->msg nick uri line)) (str->lines str)))
181
182 (define (hash-sha1 str)
183 (define in (open-input-string str))
184 (define digest (sha1 in))
185 (close-input-port in)
186 digest)
187
188 (define (uri-fetch use-cache uri)
189 (define cache-file-path
190 (expand-user-path (string-append "~/.tt/cache/" (hash-sha1 uri))))
191 (if (and use-cache (file-exists? cache-file-path))
192 (begin
193 (log-info "uri-fetch cached ~a" uri)
194 (file->string cache-file-path))
195 (begin
196 (log-info "uri-fetch new ~a" uri)
197 ; TODO Timeout. Currently hangs on slow connections.
198 (let* ([resp (http-get uri)]
199 [status (http-response-code resp)]
200 [body (http-response-body resp)])
201 (log-debug "finished GET ~a status:~a body length:~a"
202 uri status (string-length body))
203 ; TODO Handle redirects
204 (if (= status 200)
205 (begin
206 (display-to-file body cache-file-path #:exists 'replace)
207 body)
208 ; TODO A more-informative exception
209 (raise status))))))
210
211 (define (timeline-print out-format timeline)
212 (void (foldl (match-lambda**
213 [((and m (msg _ _ nick _ _)) (cons prev-nick i))
214 (let ([i (if (string=? prev-nick nick) i (+ 1 i))])
215 (msg-print out-format i m)
216 (cons nick i))])
217 (cons "" 0)
218 timeline)))
219
220 (define (feed->msgs use-cache feed)
221 (log-info "downloading feed nick:~a uri:~a"
222 (feed-nick feed)
223 (feed-uri feed))
224 (with-handlers
225 ([exn:fail:network?
226 (λ (e)
227 (log-error "network error nick:~a uri:~a exn:~a"
228 (feed-nick feed)
229 (feed-uri feed)
230 e)
231 #f)]
232 [integer?
233 (λ (status)
234 (log-error "http error nick:~a uri:~a status:~a"
235 (feed-nick feed)
236 (feed-uri feed)
237 status)
238 #f)])
239 (define uri (feed-uri feed))
240 (str->msgs [feed-nick feed] uri [uri-fetch use-cache uri])))
241
242 ; TODO timeline contract : time-sorted list of messages
243 (define (timeline use-cache num_workers feeds)
244 (sort (append* (concurrent-filter-map num_workers (curry feed->msgs use-cache) feeds))
245 (λ (a b) [< (msg-ts_epoch a) (msg-ts_epoch b)])))
246
247 (define (str->feed str)
248 ; TODO validation
249 (define toks (string-split str))
250 (apply feed toks))
251
252 (define (str->feeds str)
253 (map str->feed (str->lines str)))
254
255 (define (file->feeds filename)
256 (str->feeds (file->string filename)))
257
258 (define (user-agent prog-name prog-version)
259 (let*
260 ([prog-uri "https://github.com/xandkar/tt"]
261 [user-feed-file (expand-user-path "~/twtxt-me.txt")]
262 [user
263 (if (file-exists? user-feed-file)
264 (let ([user (first (file->feeds user-feed-file))])
265 (format "+~a; @~a" (feed-uri user) (feed-nick user)))
266 (format "+~a" prog-uri))]
267 )
268 (format "~a/~a (~a)" prog-name prog-version user)))
269
270 (define (start-logger level)
271 (let* ([logger (make-logger #f #f level #f)]
272 [log-receiver (make-log-receiver logger level)])
273 (void (thread (λ ()
274 (parameterize
275 ([date-display-format 'iso-8601])
276 (let loop ()
277 (define data (sync log-receiver))
278 (define level (vector-ref data 0))
279 (define msg (vector-ref data 1))
280 (define ts (date->string (current-date) #t))
281 (eprintf "~a [~a] ~a~n" ts level msg)
282 (loop))))))
283 (current-logger logger)))
284
285 (module+ main
286 (require setup/getinfo)
287
288 (current-http-response-auto #f)
289 (let* ([prog-name "tt"]
290 [prog-version ((get-info (list prog-name)) 'version)]
291 [user-agent (user-agent prog-name prog-version)])
292 (current-http-user-agent user-agent))
293 (let* ([use-cache
294 #f]
295 [log-level
296 'info]
297 [out-format
298 'multi-line]
299 [num_workers
300 15]) ; 15 was fastest out of the tried 1, 5, 10, 15 and 20.
301 (command-line
302 #:once-each
303 [("-c" "--cached")
304 "Read cached data instead of downloading."
305 (set! use-cache #t)]
306
307 [("-d" "--debug")
308 "Enable debug log level."
309 (set! log-level 'debug)]
310
311 [("-j" "--jobs")
312 njobs "Number of concurrent jobs."
313 (set! num_workers (string->number njobs))]
314
315 #:once-any
316 [("-s" "--short")
317 "Short output format"
318 (set! out-format 'single-line)]
319
320 [("-l" "--long")
321 "Long output format"
322 (set! out-format 'multi-line)]
323
324 #:args (filename)
325 (start-logger log-level)
326 (timeline-print out-format
327 (timeline use-cache
328 num_workers
329 (file->feeds filename))))))
This page took 0.084181 seconds and 5 git commands to generate.