3 * Stefano Zacchiroli <zack@cs.unibo.it>
4 * for the HELM Team http://helm.cs.unibo.it/
6 * This file is part of HELM, an Hypertextual, Electronic
7 * Library of Mathematics, developed at the Computer Science
8 * Department, University of Bologna, Italy.
10 * HELM is free software; you can redistribute it and/or
11 * modify it under the terms of the GNU General Public License
12 * as published by the Free Software Foundation; either version 2
13 * of the License, or (at your option) any later version.
15 * HELM is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with HELM; if not, write to the Free Software
22 * Foundation, Inc., 59 Temple Place - Suite 330, Boston,
25 * For details, see the HELM World-Wide-Web page,
26 * http://helm.cs.unibo.it/
29 open Http_getter_common;;
30 open Http_getter_misc;;
31 open Http_getter_types;;
32 open Http_getter_debugger;;
37 let common_headers = [
38 "Cache-Control", "no-cache";
43 (* HTTP queries argument parsing *)
45 (* parse encoding ("format" parameter), default is Enc_normal *)
46 let parse_enc (req: Http_types.request) =
48 (match req#param "format" with
49 | "normal" -> Enc_normal
51 | s -> raise (Http_getter_bad_request ("Invalid format: " ^ s)))
52 with Http_types.Param_not_found _ -> Enc_normal
54 (* parse "patch_dtd" parameter, default is true *)
55 let parse_patch (req: Http_types.request) =
57 (match req#param "patch_dtd" with
58 | s when String.lowercase s = "yes" -> true
59 | s when String.lowercase s = "no" -> false
60 | s -> raise (Http_getter_bad_request ("Invalid patch_dtd value: " ^ s)))
61 with Http_types.Param_not_found _ -> true
63 (* parse output format ("format" parameter), no default value *)
64 let parse_output_format (req: Http_types.request) =
65 match req#param "format" with
66 | s when String.lowercase s = "txt" -> Fmt_text
67 | s when String.lowercase s = "xml" -> Fmt_xml
68 | s -> raise (Http_getter_bad_request ("Invalid /ls format: " ^ s))
70 (* parse "baseuri" format for /ls method, no default value *)
72 let parse_ls_RE = Pcre.regexp "^(\\w+):(.*)$" in
73 let trailing_slash_RE = Pcre.regexp "/+$" in
74 fun (req: Http_types.request) ->
75 let baseuri = req#param "baseuri" in
77 Pcre.extract ~rex:parse_ls_RE
78 (Pcre.replace ~rex:trailing_slash_RE baseuri)
80 match (subs.(1), subs.(2)) with
81 | "cic", uri -> Cic uri
82 | "theory", uri -> Theory uri
83 | _ -> raise (Http_getter_bad_request ("Invalid /ls baseuri: " ^ baseuri))
86 (* global maps, shared by all threads *)
88 let cic_map = new Http_getter_map.map Http_getter_env.cic_dbm in
89 let nuprl_map = new Http_getter_map.map Http_getter_env.nuprl_dbm in
90 let rdf_map = new Http_getter_map.map Http_getter_env.rdf_dbm in
91 let xsl_map = new Http_getter_map.map Http_getter_env.xsl_dbm in
94 cic_map#close; nuprl_map#close; rdf_map#close; xsl_map#close in
95 let map_of_uri = function
96 | uri when is_cic_uri uri -> cic_map
97 | uri when is_nuprl_uri uri -> nuprl_map
98 | uri when is_rdf_uri uri -> rdf_map
99 | uri when is_xsl_uri uri -> xsl_map
100 | uri -> raise (Http_getter_unresolvable_URI uri)
104 (map_of_uri uri)#resolve uri
105 with Http_getter_map.Key_not_found _ ->
106 raise (Http_getter_unresolvable_URI uri)
109 (* Warning: this fail if uri is already registered *)
110 (map_of_uri uri)#add uri
112 let return_all_foo_uris map doctype filter outchan =
113 (** return all URIs contained in 'map' which satisfy predicate 'filter'; URIs
114 are written in an XMLish format ('doctype' is the XML doctype) onto 'outchan'
116 Http_daemon.send_basic_headers ~code:200 outchan;
117 Http_daemon.send_header "Content-Type" "text/xml" outchan;
118 Http_daemon.send_headers common_headers outchan;
119 Http_daemon.send_CRLF outchan;
123 "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>
124 <!DOCTYPE %s SYSTEM \"%s/getdtd?uri=%s.dtd\">
129 Http_getter_env.my_own_url
135 output_string outchan (sprintf "\t<uri value=\"%s\" />\n" uri));
136 output_string outchan (sprintf "</%s>\n" doctype)
138 let return_all_xml_uris = return_all_foo_uris cic_map "alluris" in
139 let return_all_rdf_uris = return_all_foo_uris rdf_map "allrdfuris" in
141 let (++) (oldann, oldtypes, oldbody) (newann, newtypes, newbody) =
142 ((if newann > oldann then newann else oldann),
143 (if newtypes > oldtypes then newtypes else oldtypes),
144 (if newbody > oldbody then newbody else oldbody))
147 Pcre.regexp "^([^.]*\\.[^.]*)((\\.body)|(\\.types))?(\\.ann)?"
149 let (types_RE, types_ann_RE, body_RE, body_ann_RE) =
150 (Pcre.regexp "\\.types", Pcre.regexp "\\.types.ann",
151 Pcre.regexp "\\.body", Pcre.regexp "\\.body.ann")
153 let (slash_RE, til_slash_RE, no_slashes_RE) =
154 (Pcre.regexp "/", Pcre.regexp "^.*/", Pcre.regexp "^[^/]*$")
156 fun lsuri fmt outchan ->
159 (match lsuri with Cic p -> ("cic:" ^ p) | Theory p -> ("theory:" ^ p))
161 let (dir_RE, obj_RE) =
162 (Pcre.regexp (pat ^ "/"), Pcre.regexp (pat ^ "(\\.|$)"))
164 let dirs = ref StringSet.empty in
165 let objs = Hashtbl.create 17 in
167 dirs := StringSet.add (List.hd (Pcre.split ~rex:slash_RE d)) !dirs
170 let basepart = Pcre.replace ~rex:basepart_RE ~templ:"$1" o in
173 Hashtbl.find objs basepart
174 with Not_found -> (false, No, No) (* no ann, no types, no body *)
178 | s when Pcre.pmatch ~rex:types_RE s -> (false, Yes, No)
179 | s when Pcre.pmatch ~rex:types_ann_RE s -> (true, Ann, No)
180 | s when Pcre.pmatch ~rex:body_RE s -> (false, No, Yes)
181 | s when Pcre.pmatch ~rex:body_ann_RE s -> (true, No, Ann)
182 | s -> (false, No, No)
184 Hashtbl.replace objs basepart (oldflags ++ newflags)
186 cic_map#iter (* BLEARGH Dbm module lacks support for fold-like functions *)
189 | uri when Pcre.pmatch ~rex:dir_RE uri -> (* directory hit *)
190 let localpart = Pcre.replace ~rex:dir_RE uri in
191 if Pcre.pmatch ~rex:no_slashes_RE localpart then
195 | uri when Pcre.pmatch ~rex:obj_RE uri -> (* file hit *)
196 store_obj (Pcre.replace ~rex:til_slash_RE uri)
197 | uri -> () (* miss *));
202 (fun s d -> sprintf "%sdir, %s\n" s d) ""
203 (StringSet.elements !dirs)) ^
204 (Http_getter_misc.hashtbl_sorted_fold
205 (fun uri (annflag, typesflag, bodyflag) cont ->
206 sprintf "%sobject, %s, <%s,%s,%s>\n"
207 cont uri (if annflag then "YES" else "NO")
208 (string_of_ls_flag typesflag) (string_of_ls_flag bodyflag))
212 ~headers:(("Content-Type", "text/plain") :: common_headers)
217 "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>
218 <!DOCTYPE ls SYSTEM \"%s/getdtd?uri=ls.dtd\">
224 Http_getter_env.my_own_url
229 (fun d -> "<section>" ^ d ^ "</section>")
230 (StringSet.elements !dirs))) ^ "\n" ^
231 (Http_getter_misc.hashtbl_sorted_fold
232 (fun uri (annflag, typesflag, bodyflag) cont ->
234 "%s<object name=\"%s\">
235 \t<ann value=\"%s\" />
236 \t<types value=\"%s\" />
237 \t<body value=\"%s\" />
240 cont uri (if annflag then "YES" else "NO")
241 (string_of_ls_flag typesflag)
242 (string_of_ls_flag bodyflag))
246 ~headers:(("Content-Type", "text/xml") :: common_headers)
249 let (index_line_sep_RE, index_sep_RE, trailing_types_RE,
250 heading_cic_RE, heading_theory_RE, heading_nuprl_RE,
251 heading_rdf_cic_RE, heading_rdf_theory_RE) =
252 (Pcre.regexp "[ \t]+", Pcre.regexp "\r\n|\r|\n",
253 Pcre.regexp "\\.types$",
254 Pcre.regexp "^cic:", Pcre.regexp "^theory:", Pcre.regexp "^nuprl:",
255 Pcre.regexp "^helm:rdf.*//cic:", Pcre.regexp "^helm:rdf.*//theory:")
257 let update_from_server logmsg server_url = (* use global maps *)
258 debug_print ("Updating information from " ^ server_url);
259 let xml_url_of_uri = function
260 (* TODO missing sanity checks on server_url, e.g. it can contains $1 *)
261 | uri when (Pcre.pmatch ~rex:heading_cic_RE uri) ->
262 Pcre.replace ~rex:heading_cic_RE ~templ:server_url uri
263 | uri when (Pcre.pmatch ~rex:heading_theory_RE uri) ->
264 Pcre.replace ~rex:heading_theory_RE ~templ:server_url uri
265 | uri when (Pcre.pmatch ~rex:heading_nuprl_RE uri) ->
266 Pcre.replace ~rex:heading_nuprl_RE ~templ:server_url uri
267 | uri -> raise (Http_getter_invalid_URI uri)
269 let rdf_url_of_uri = function (* TODO as above *)
270 | uri when (Pcre.pmatch ~rex:heading_rdf_cic_RE uri) ->
271 Pcre.replace ~rex:heading_rdf_cic_RE ~templ:server_url uri
272 | uri when (Pcre.pmatch ~rex:heading_rdf_theory_RE uri) ->
273 Pcre.replace ~rex:heading_rdf_theory_RE ~templ:server_url uri
274 | uri -> raise (Http_getter_invalid_URI uri)
276 let log = ref (logmsg ^ "Processing server: " ^ server_url ^ "<br />\n") in
277 let (xml_index, rdf_index, xsl_index) =
278 (* TODO keeps index in memory, is better to keep them on temp files? *)
279 (http_get (server_url ^ "/" ^ Http_getter_env.xml_index),
280 http_get (server_url ^ "/" ^ Http_getter_env.rdf_index),
281 http_get (server_url ^ "/" ^ Http_getter_env.xsl_index))
283 if (xml_index = None && rdf_index = None && xsl_index = None) then
284 debug_print (sprintf "Warning: useless server %s" server_url);
285 (match xml_index with
287 (log := !log ^ "Updating XML db ...<br />\n";
290 | l when is_blank_line l -> () (* skip blank and commented lines *)
293 (match Pcre.split ~rex:index_line_sep_RE l with
295 assert (is_cic_uri uri || is_nuprl_uri uri) ;
296 (map_of_uri uri)#replace
297 uri ((xml_url_of_uri uri) ^ ".xml.gz")
299 assert (is_cic_uri uri || is_nuprl_uri uri) ;
300 (map_of_uri uri)#replace
301 uri ((xml_url_of_uri uri) ^ ".xml")
303 log := !log ^ "Ignoring invalid line: '" ^ l ^ "'<br />\n")
304 with Http_getter_invalid_URI uri ->
305 log := !log ^ "Ignoring invalid XML URI: '" ^ uri ^ "'<br />\n")
306 (Pcre.split ~rex:index_sep_RE xml_index)) (* xml_index lines *)
308 (match rdf_index with
310 (log := !log ^ "Updating RDF db ...<br />\n";
314 (match Pcre.split ~rex:index_line_sep_RE l with
316 rdf_map#replace uri ((rdf_url_of_uri uri) ^ ".xml.gz")
317 | [uri] -> rdf_map#replace uri ((rdf_url_of_uri uri) ^ ".xml")
318 | _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "<br />\n")
319 with Http_getter_invalid_URI uri ->
320 log := !log ^ "Ignoring invalid RDF URI: " ^ uri ^ "<br />\n")
321 (Pcre.split ~rex:index_sep_RE rdf_index)) (* rdf_index lines *)
323 (match xsl_index with
325 (log := !log ^ "Updating XSLT db ...<br />\n";
327 (fun l -> xsl_map#replace l (server_url ^ "/" ^ l))
328 (Pcre.split ~rex:index_sep_RE xsl_index);
329 log := !log ^ "All done!<br />\n")
331 debug_print "done with this server";
334 let update_from_all_servers () = (* use global maps *)
335 cic_map#clear; nuprl_map#clear; rdf_map#clear; xsl_map#clear;
339 "" (* initial logmsg: empty *)
340 (* reverse order: 1st server is the most important one *)
341 (List.rev !Http_getter_env.servers)
343 cic_map#sync; nuprl_map#sync; rdf_map#sync; xsl_map#sync;
349 let callback (req: Http_types.request) outchan =
351 debug_print ("Connection from " ^ req#clientAddr);
352 debug_print ("Received request: " ^ req#path);
356 (Http_getter_const.usage_string (Http_getter_env.env_to_string ()))
358 | "/getxml" | "/getxslt" | "/getdtd" | "/resolve" | "/register" ->
359 (let uri = req#param "uri" in (* common parameter *)
362 let enc = parse_enc req in
363 let patch = parse_patch req in
364 Http_getter_cache.respond_xml
365 ~url:(resolve uri) ~uri ~enc ~patch outchan
367 let patch = parse_patch req in
368 Http_getter_cache.respond_xsl ~url:(resolve uri) ~patch outchan
370 let patch = parse_patch req in
371 Http_getter_cache.respond_dtd
372 ~patch ~url:(Http_getter_env.dtd_dir ^ "/" ^ uri) outchan
376 (sprintf "<url value=\"%s\" />\n" (resolve uri))
378 with Http_getter_unresolvable_URI uri ->
379 return_xml_raw "<unresolved />\n" outchan)
381 let url = req#param "url" in
383 return_html_msg "Register done" outchan
386 Http_getter_env.reload (); (* reload servers list from servers file *)
387 let log = update_from_all_servers () in
388 return_html_msg log outchan
391 (sprintf "<html><body><table>\n%s\n</table></body></html>"
395 fun s -> incr i; sprintf "<tr><td>%d</td><td>%s</td></tr>" !i s)
396 !Http_getter_env.servers)))
399 let name = req#param "url" in
403 let res = int_of_string (req#param "position") in
405 raise (Failure "int_of_string");
407 with Failure "int_of_string" ->
408 raise (Http_getter_bad_request
409 (sprintf "position must be a non negative integer (%s given)"
410 (req#param "position")))
412 if position = 0 then (* fallback to default value *)
413 raise (Http_types.Param_not_found "foo")
414 else if position > 0 then begin (* add server and update all *)
415 Http_getter_env.add_server ~position name;
416 let log = update_from_all_servers () in
418 (sprintf "Added server %s in position %d)<br />\n%s"
421 end else (* position < 0 *) (* error! *)
422 assert false (* already checked above *)
423 with Http_types.Param_not_found _ -> (* add as 1st server by default *)
424 Http_getter_env.add_server ~position:0 name;
425 let log = update_from_server (* quick update (new server only) *)
426 (sprintf "Added server %s in head position<br />\n" name) name
428 return_html_msg log outchan)
429 | "/remove_server" ->
432 let res = int_of_string (req#param "position") in
434 raise (Failure "int_of_string");
436 with Failure "int_of_string" ->
437 raise (Http_getter_bad_request
438 (sprintf "position must be a non negative integer (%s given)"
439 (req#param "position")))
443 List.nth !Http_getter_env.servers position
444 with Failure "nth" ->
445 raise (Http_getter_bad_request
446 (sprintf "no server with position %d" position))
448 Http_getter_env.remove_server position;
449 let log = update_from_all_servers () in
451 (sprintf "Removed server %s (position %d)<br />\n%s"
452 server_name position log)
457 (Pcre.pmatch ~rex:heading_cic_RE uri) &&
458 not (Pcre.pmatch ~rex:trailing_types_RE uri))
460 | "/getallrdfuris" ->
461 (let classs = req#param "class" in
464 let base = "^helm:rdf:www\\.cs\\.unibo\\.it/helm/rdf/" in
466 | ("forward" as c) | ("backward" as c) ->
467 (fun uri -> Pcre.pmatch ~pat:(base ^ c) uri)
468 | c -> raise (Http_getter_invalid_RDF_class c)
470 return_all_rdf_uris filter outchan
471 with Http_getter_invalid_RDF_class c ->
472 raise (Http_getter_bad_request ("Invalid RDF class: " ^ c)))
473 | "/ls" -> return_ls (parse_ls_uri req) (parse_output_format req) outchan
475 Http_daemon.respond ~body:Http_getter_const.empty_xml outchan
477 Http_daemon.respond_error ~status:(`Client_error `Bad_request) outchan);
478 debug_print "Done!\n"
480 | Http_types.Param_not_found attr_name ->
481 return_400 (sprintf "Parameter '%s' is missing" attr_name) outchan
482 | Http_getter_bad_request msg -> return_html_error msg outchan
483 | Http_getter_internal_error msg -> return_html_internal_error msg outchan
484 | Shell.Subprocess_error l ->
485 return_html_internal_error
486 (String.concat "<br />\n"
489 sprintf "Command '%s' returned %s"
490 cmd (string_of_proc_status code))
495 ("Uncaught exception: " ^ (Printexc.to_string exc))
499 (* daemon initialization *)
502 print_string (Http_getter_env.env_to_string ());
504 Unix.putenv "http_proxy" "";
506 Sys.catch_break true;
509 ~timeout:(Some 600) ~port:Http_getter_env.port ~mode:`Thread callback
510 with Sys.Break -> () (* 'save_maps' already registered with 'at_exit' *)