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)
108 let register uri = (map_of_uri uri )#add uri in
109 let return_all_foo_uris map doctype filter outchan =
110 (** return all URIs contained in 'map' which satisfy predicate 'filter'; URIs
111 are written in an XMLish format ('doctype' is the XML doctype) onto 'outchan'
113 Http_daemon.send_basic_headers ~code:200 outchan;
114 Http_daemon.send_header "Content-Type" "text/xml" outchan;
115 Http_daemon.send_headers common_headers outchan;
116 Http_daemon.send_CRLF outchan;
120 "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>
121 <!DOCTYPE %s SYSTEM \"%s/getdtd?uri=%s.dtd\">
126 Http_getter_env.my_own_url
132 output_string outchan (sprintf "\t<uri value=\"%s\" />\n" uri));
133 output_string outchan (sprintf "</%s>\n" doctype)
135 let return_all_xml_uris = return_all_foo_uris cic_map "alluris" in
136 let return_all_rdf_uris = return_all_foo_uris rdf_map "allrdfuris" in
138 let (++) (oldann, oldtypes, oldbody) (newann, newtypes, newbody) =
139 ((if newann > oldann then newann else oldann),
140 (if newtypes > oldtypes then newtypes else oldtypes),
141 (if newbody > oldbody then newbody else oldbody))
144 Pcre.regexp "^([^.]*\\.[^.]*)((\\.body)|(\\.types))?(\\.ann)?"
146 let (types_RE, types_ann_RE, body_RE, body_ann_RE) =
147 (Pcre.regexp "\\.types", Pcre.regexp "\\.types.ann",
148 Pcre.regexp "\\.body", Pcre.regexp "\\.body.ann")
150 let (slash_RE, til_slash_RE, no_slashes_RE) =
151 (Pcre.regexp "/", Pcre.regexp "^.*/", Pcre.regexp "^[^/]*$")
153 fun lsuri fmt outchan ->
156 (match lsuri with Cic p -> ("cic:" ^ p) | Theory p -> ("theory:" ^ p))
158 let (dir_RE, obj_RE) =
159 (Pcre.regexp (pat ^ "/"), Pcre.regexp (pat ^ "(\\.|$)"))
161 let dirs = ref StringSet.empty in
162 let objs = Hashtbl.create 17 in
164 dirs := StringSet.add (List.hd (Pcre.split ~rex:slash_RE d)) !dirs
167 let basepart = Pcre.replace ~rex:basepart_RE ~templ:"$1" o in
170 Hashtbl.find objs basepart
171 with Not_found -> (false, No, No) (* no ann, no types, no body *)
175 | s when Pcre.pmatch ~rex:types_RE s -> (false, Yes, No)
176 | s when Pcre.pmatch ~rex:types_ann_RE s -> (true, Ann, No)
177 | s when Pcre.pmatch ~rex:body_RE s -> (false, No, Yes)
178 | s when Pcre.pmatch ~rex:body_ann_RE s -> (true, No, Ann)
179 | s -> (false, No, No)
181 Hashtbl.replace objs basepart (oldflags ++ newflags)
183 cic_map#iter (* BLEARGH Dbm module lacks support for fold-like functions *)
186 | uri when Pcre.pmatch ~rex:dir_RE uri -> (* directory hit *)
187 let localpart = Pcre.replace ~rex:dir_RE uri in
188 if Pcre.pmatch ~rex:no_slashes_RE localpart then
192 | uri when Pcre.pmatch ~rex:obj_RE uri -> (* file hit *)
193 store_obj (Pcre.replace ~rex:til_slash_RE uri)
194 | uri -> () (* miss *));
199 (fun s d -> sprintf "%sdir, %s\n" s d) ""
200 (StringSet.elements !dirs)) ^
201 (Http_getter_misc.hashtbl_sorted_fold
202 (fun uri (annflag, typesflag, bodyflag) cont ->
203 sprintf "%sobject, %s, <%s,%s,%s>\n"
204 cont uri (if annflag then "YES" else "NO")
205 (string_of_ls_flag typesflag) (string_of_ls_flag bodyflag))
209 ~headers:(("Content-Type", "text/plain") :: common_headers)
214 "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>
215 <!DOCTYPE ls SYSTEM \"%s/getdtd?uri=ls.dtd\">
221 Http_getter_env.my_own_url
226 (fun d -> "<section>" ^ d ^ "</section>")
227 (StringSet.elements !dirs))) ^ "\n" ^
228 (Http_getter_misc.hashtbl_sorted_fold
229 (fun uri (annflag, typesflag, bodyflag) cont ->
231 "%s<object name=\"%s\">
232 \t<ann value=\"%s\" />
233 \t<types value=\"%s\" />
234 \t<body value=\"%s\" />
237 cont uri (if annflag then "YES" else "NO")
238 (string_of_ls_flag typesflag)
239 (string_of_ls_flag bodyflag))
243 ~headers:(("Content-Type", "text/xml") :: common_headers)
246 let (index_line_sep_RE, index_sep_RE, trailing_types_RE,
247 heading_cic_RE, heading_theory_RE, heading_nuprl_RE,
248 heading_rdf_cic_RE, heading_rdf_theory_RE) =
249 (Pcre.regexp "[ \t]+", Pcre.regexp "\r\n|\r|\n",
250 Pcre.regexp "\\.types$",
251 Pcre.regexp "^cic:", Pcre.regexp "^theory:", Pcre.regexp "^nuprl:",
252 Pcre.regexp "^helm:rdf.*//cic:", Pcre.regexp "^helm:rdf.*//theory:")
254 let update_from_server logmsg server_url = (* use global maps *)
255 debug_print ("Updating information from " ^ server_url);
256 let xml_url_of_uri = function
257 (* TODO missing sanity checks on server_url, e.g. it can contains $1 *)
258 | uri when (Pcre.pmatch ~rex:heading_cic_RE uri) ->
259 Pcre.replace ~rex:heading_cic_RE ~templ:server_url uri
260 | uri when (Pcre.pmatch ~rex:heading_theory_RE uri) ->
261 Pcre.replace ~rex:heading_theory_RE ~templ:server_url uri
262 | uri when (Pcre.pmatch ~rex:heading_nuprl_RE uri) ->
263 Pcre.replace ~rex:heading_nuprl_RE ~templ:server_url uri
264 | uri -> raise (Http_getter_invalid_URI uri)
266 let rdf_url_of_uri = function (* TODO as above *)
267 | uri when (Pcre.pmatch ~rex:heading_rdf_cic_RE uri) ->
268 Pcre.replace ~rex:heading_rdf_cic_RE ~templ:server_url uri
269 | uri when (Pcre.pmatch ~rex:heading_rdf_theory_RE uri) ->
270 Pcre.replace ~rex:heading_rdf_theory_RE ~templ:server_url uri
271 | uri -> raise (Http_getter_invalid_URI uri)
273 let log = ref (logmsg ^ "Processing server: " ^ server_url ^ "<br />\n") in
274 let (xml_index, rdf_index, xsl_index) =
275 (* TODO keeps index in memory, is better to keep them on temp files? *)
276 (http_get (server_url ^ "/" ^ Http_getter_env.xml_index),
277 http_get (server_url ^ "/" ^ Http_getter_env.rdf_index),
278 http_get (server_url ^ "/" ^ Http_getter_env.xsl_index))
280 if (xml_index = None && rdf_index = None && xsl_index = None) then
281 debug_print (sprintf "Warning: useless server %s" server_url);
282 (match xml_index with
284 (log := !log ^ "Updating XML db ...<br />\n";
287 | l when is_blank_line l -> () (* skip blank and commented lines *)
290 (match Pcre.split ~rex:index_line_sep_RE l with
292 assert (is_cic_uri uri || is_nuprl_uri uri) ;
293 (map_of_uri uri)#add uri ((xml_url_of_uri uri) ^ ".xml.gz")
295 assert (is_cic_uri uri || is_nuprl_uri uri) ;
296 (map_of_uri uri)#add uri ((xml_url_of_uri uri) ^ ".xml")
298 log := !log ^ "Ignoring invalid line: '" ^ l ^ "'<br />\n")
299 with Http_getter_invalid_URI uri ->
300 log := !log ^ "Ignoring invalid XML URI: '" ^ uri ^ "'<br />\n")
301 (Pcre.split ~rex:index_sep_RE xml_index)) (* xml_index lines *)
303 (match rdf_index with
305 (log := !log ^ "Updating RDF db ...<br />\n";
309 (match Pcre.split ~rex:index_line_sep_RE l with
310 | [uri; "gz"] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml.gz")
311 | [uri] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml")
312 | _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "<br />\n")
313 with Http_getter_invalid_URI uri ->
314 log := !log ^ "Ignoring invalid RDF URI: " ^ uri ^ "<br />\n")
315 (Pcre.split ~rex:index_sep_RE rdf_index)) (* rdf_index lines *)
317 (match xsl_index with
319 (log := !log ^ "Updating XSLT db ...<br />\n";
321 (fun l -> xsl_map#add l (server_url ^ "/" ^ l))
322 (Pcre.split ~rex:index_sep_RE xsl_index);
323 log := !log ^ "All done!<br />\n")
325 debug_print "done with this server";
331 let callback (req: Http_types.request) outchan =
333 debug_print ("Connection from " ^ req#clientAddr);
334 debug_print ("Received request: " ^ req#path);
338 (Http_getter_const.usage_string (Http_getter_env.env_to_string ()))
340 | "/getxml" | "/getxslt" | "/getdtd" | "/resolve" | "/register" ->
341 (let uri = req#param "uri" in (* common parameter *)
344 let enc = parse_enc req in
345 let patch = parse_patch req in
346 Http_getter_cache.respond_xml
347 ~url:(resolve uri) ~uri ~enc ~patch outchan
349 let patch = parse_patch req in
350 Http_getter_cache.respond_xsl ~url:(resolve uri) ~patch outchan
352 let patch = parse_patch req in
353 Http_getter_cache.respond_dtd
354 ~patch ~url:(Http_getter_env.dtd_dir ^ "/" ^ uri) outchan
358 (sprintf "<url value=\"%s\" />\n" (resolve uri))
360 with Http_getter_unresolvable_URI uri ->
361 return_xml_raw "<unresolved />\n" outchan)
363 let url = req#param "url" in
365 return_html_msg "Register done" outchan
368 (Http_getter_env.reload (); (* reload servers list from servers file *)
369 cic_map#clear; nuprl_map#clear; rdf_map#clear; xsl_map#clear;
373 "" (* initial logmsg: empty *)
374 (* reverse order: 1st server is the most important one *)
375 (List.rev !Http_getter_env.servers)
377 cic_map#sync; nuprl_map#sync; rdf_map#sync; xsl_map#sync;
378 return_html_msg log outchan)
382 (Pcre.pmatch ~rex:heading_cic_RE uri) &&
383 not (Pcre.pmatch ~rex:trailing_types_RE uri))
385 | "/getallrdfuris" ->
386 (let classs = req#param "class" in
389 let base = "^helm:rdf:www\\.cs\\.unibo\\.it/helm/rdf/" in
391 | ("forward" as c) | ("backward" as c) ->
392 (fun uri -> Pcre.pmatch ~pat:(base ^ c) uri)
393 | c -> raise (Http_getter_invalid_RDF_class c)
395 return_all_rdf_uris filter outchan
396 with Http_getter_invalid_RDF_class c ->
397 raise (Http_getter_bad_request ("Invalid RDF class: " ^ c)))
398 | "/ls" -> return_ls (parse_ls_uri req) (parse_output_format req) outchan
400 Http_daemon.respond ~body:Http_getter_const.empty_xml outchan
402 Http_daemon.respond_error ~status:(`Client_error `Bad_request) outchan);
403 debug_print "Done!\n"
405 | Http_types.Param_not_found attr_name ->
406 return_400 (sprintf "Parameter '%s' is missing" attr_name) outchan
407 | Http_getter_bad_request msg -> return_html_error msg outchan
408 | Http_getter_internal_error msg -> return_html_internal_error msg outchan
409 | Shell.Subprocess_error l ->
410 return_html_internal_error
411 (String.concat "<br />\n"
414 sprintf "Command '%s' returned %s"
415 cmd (string_of_proc_status code))
420 ("Uncaught exception: " ^ (Printexc.to_string exc))
424 (* daemon initialization *)
427 print_string (Http_getter_env.env_to_string ());
429 Unix.putenv "http_proxy" "";
431 Sys.catch_break true;
434 ~timeout:(Some 600) ~port:Http_getter_env.port ~mode:`Thread callback
435 with Sys.Break -> () (* 'save_maps' already registered with 'at_exit' *)