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 let parse_enc (req: Http_types.request) =
47 (match req#param "format" with
48 | "normal" -> Enc_normal
50 | s -> raise (Http_getter_bad_request ("Invalid format: " ^ s)))
51 with Http_types.Param_not_found _ -> Enc_normal
53 let parse_patch (req: Http_types.request) =
54 match req#param "patch_dtd" with
55 | s when String.lowercase s = "yes" -> true
56 | s when String.lowercase s = "no" -> false
57 | s -> raise (Http_getter_bad_request ("Invalid patch_dtd value: " ^ s))
59 let parse_output_format (req: Http_types.request) =
60 match req#param "format" with
61 | s when String.lowercase s = "txt" -> Fmt_text
62 | s when String.lowercase s = "xml" -> Fmt_xml
63 | s -> raise (Http_getter_bad_request ("Invalid /ls format: " ^ s))
66 let parse_ls_RE = Pcre.regexp "^(\\w+):(.*)$" in
67 let trailing_slash_RE = Pcre.regexp "/+$" in
68 fun (req: Http_types.request) ->
69 let baseuri = req#param "baseuri" in
71 Pcre.extract ~rex:parse_ls_RE
72 (Pcre.replace ~rex:trailing_slash_RE baseuri)
74 match (subs.(1), subs.(2)) with
75 | "cic", uri -> Cic uri
76 | "theory", uri -> Theory uri
77 | _ -> raise (Http_getter_bad_request ("Invalid /ls baseuri: " ^ baseuri))
80 (* global maps, shared by all threads *)
82 let cic_map = new Http_getter_map.map Http_getter_env.cic_dbm in
83 let nuprl_map = new Http_getter_map.map Http_getter_env.nuprl_dbm in
84 let rdf_map = new Http_getter_map.map Http_getter_env.rdf_dbm in
85 let xsl_map = new Http_getter_map.map Http_getter_env.xsl_dbm in
88 cic_map#close; nuprl_map#close; rdf_map#close; xsl_map#close in
89 let map_of_uri = function
90 | uri when is_cic_uri uri -> cic_map
91 | uri when is_nuprl_uri uri -> nuprl_map
92 | uri when is_rdf_uri uri -> rdf_map
93 | uri when is_xsl_uri uri -> xsl_map
94 | uri -> raise (Http_getter_unresolvable_URI uri)
98 (map_of_uri uri)#resolve uri
99 with Http_getter_map.Key_not_found _ ->
100 raise (Http_getter_unresolvable_URI uri)
102 let register uri = (map_of_uri uri )#add uri in
103 let return_all_foo_uris map doctype filter outchan =
104 (** return all URIs contained in 'map' which satisfy predicate 'filter'; URIs
105 are written in an XMLish format ('doctype' is the XML doctype) onto 'outchan'
107 Http_daemon.send_basic_headers ~code:200 outchan;
108 Http_daemon.send_header "Content-Type" "text/xml" outchan;
109 Http_daemon.send_headers common_headers outchan;
110 Http_daemon.send_CRLF outchan;
114 "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>
115 <!DOCTYPE %s SYSTEM \"%s/getdtd?uri=%s.dtd\">
120 Http_getter_env.my_own_url
126 output_string outchan (sprintf "\t<uri value=\"%s\" />\n" uri));
127 output_string outchan (sprintf "</%s>\n" doctype)
129 let return_all_xml_uris = return_all_foo_uris cic_map "alluris" in
130 let return_all_rdf_uris = return_all_foo_uris rdf_map "allrdfuris" in
132 let (++) (oldann, oldtypes, oldbody) (newann, newtypes, newbody) =
133 ((if newann > oldann then newann else oldann),
134 (if newtypes > oldtypes then newtypes else oldtypes),
135 (if newbody > oldbody then newbody else oldbody))
138 Pcre.regexp "^([^.]*\\.[^.]*)((\\.body)|(\\.types))?(\\.ann)?"
140 let (types_RE, types_ann_RE, body_RE, body_ann_RE) =
141 (Pcre.regexp "\\.types", Pcre.regexp "\\.types.ann",
142 Pcre.regexp "\\.body", Pcre.regexp "\\.body.ann")
144 let (slash_RE, til_slash_RE, no_slashes_RE) =
145 (Pcre.regexp "/", Pcre.regexp "^.*/", Pcre.regexp "^[^/]*$")
147 fun lsuri fmt outchan ->
150 (match lsuri with Cic p -> ("cic:" ^ p) | Theory p -> ("theory:" ^ p))
152 let (dir_RE, obj_RE) =
153 (Pcre.regexp (pat ^ "/"), Pcre.regexp (pat ^ "(\\.|$)"))
155 let dirs = ref StringSet.empty in
156 let objs = Hashtbl.create 17 in
158 dirs := StringSet.add (List.hd (Pcre.split ~rex:slash_RE d)) !dirs
161 let basepart = Pcre.replace ~rex:basepart_RE ~templ:"$1" o in
164 Hashtbl.find objs basepart
165 with Not_found -> (false, No, No) (* no ann, no types no body *)
169 | s when Pcre.pmatch ~rex:types_RE s -> (false, Yes, No)
170 | s when Pcre.pmatch ~rex:types_ann_RE s -> (true, Ann, No)
171 | s when Pcre.pmatch ~rex:body_RE s -> (false, No, Yes)
172 | s when Pcre.pmatch ~rex:body_ann_RE s -> (true, No, Ann)
173 | s -> (false, No, No)
175 Hashtbl.replace objs basepart (oldflags ++ newflags)
177 cic_map#iter (* BLEARGH Dbm module lacks support for fold-like functions *)
180 | uri when Pcre.pmatch ~rex:dir_RE uri -> (* directory hit *)
181 let localpart = Pcre.replace ~rex:dir_RE uri in
182 if Pcre.pmatch ~rex:no_slashes_RE localpart then
186 | uri when Pcre.pmatch ~rex:obj_RE uri -> (* file hit *)
187 store_obj (Pcre.replace ~rex:til_slash_RE uri)
188 | uri -> () (* miss *));
193 (fun s d -> sprintf "%sdir, %s\n" s d) ""
194 (StringSet.elements !dirs)) ^
195 (Http_getter_misc.hashtbl_sorted_fold
196 (fun uri (annflag, typesflag, bodyflag) cont ->
197 sprintf "%sobject, %s, <%s,%s,%s>\n"
198 cont uri (if annflag then "YES" else "NO")
199 (string_of_ls_flag typesflag) (string_of_ls_flag bodyflag))
203 ~headers:(("Content-Type", "text/plain") :: common_headers)
208 "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>
209 <!DOCTYPE ls SYSTEM \"%s/getdtd?uri=ls.dtd\">
215 Http_getter_env.my_own_url
220 (fun d -> "<section>" ^ d ^ "</section>")
221 (StringSet.elements !dirs))) ^ "\n" ^
222 (Http_getter_misc.hashtbl_sorted_fold
223 (fun uri (annflag, typesflag, bodyflag) cont ->
225 "%s<object name=\"%s\">
226 \t<ann value=\"%s\" />
227 \t<types value=\"%s\" />
228 \t<body value=\"%s\" />
231 cont uri (if annflag then "YES" else "NO")
232 (string_of_ls_flag typesflag)
233 (string_of_ls_flag bodyflag))
237 ~headers:(("Content-Type", "text/xml") :: common_headers)
240 let (index_line_sep_RE, index_sep_RE, trailing_types_RE,
241 heading_cic_RE, heading_theory_RE,
242 heading_rdf_cic_RE, heading_rdf_theory_RE) =
243 (Pcre.regexp "[ \t]+", Pcre.regexp "\n+", Pcre.regexp "\\.types$",
244 Pcre.regexp "^cic:", Pcre.regexp "^theory:",
245 Pcre.regexp "^helm:rdf.*//cic:", Pcre.regexp "^helm:rdf.*//theory:")
247 let update_from_server logmsg server_url = (* use global maps *)
248 debug_print ("Updating information from " ^ server_url);
249 let xml_url_of_uri = function
250 (* TODO missing sanity checks on server_url, e.g. it can contains $1 *)
251 | uri when (Pcre.pmatch ~rex:heading_cic_RE uri) ->
252 Pcre.replace ~rex:heading_cic_RE ~templ:server_url uri
253 | uri when (Pcre.pmatch ~rex:heading_theory_RE uri) ->
254 Pcre.replace ~rex:heading_theory_RE ~templ:server_url uri
255 | uri -> raise (Http_getter_invalid_URI uri)
257 let rdf_url_of_uri = function (* TODO as above *)
258 | uri when (Pcre.pmatch ~rex:heading_rdf_cic_RE uri) ->
259 Pcre.replace ~rex:heading_rdf_cic_RE ~templ:server_url uri
260 | uri when (Pcre.pmatch ~rex:heading_rdf_theory_RE uri) ->
261 Pcre.replace ~rex:heading_rdf_theory_RE ~templ:server_url uri
262 | uri -> raise (Http_getter_invalid_URI uri)
264 let log = ref (logmsg ^ "Processing server: " ^ server_url ^ "<br />\n") in
265 let (xml_index, rdf_index, xsl_index) =
266 (* TODO keeps index in memory, is better to keep them on temp files? *)
267 (http_get (server_url ^ "/" ^ Http_getter_env.xml_index),
268 http_get (server_url ^ "/" ^ Http_getter_env.rdf_index),
269 http_get (server_url ^ "/" ^ Http_getter_env.xsl_index))
271 if (xml_index = None && rdf_index = None && xsl_index = None) then
272 debug_print (sprintf "Warning: useless server %s" server_url);
273 (match xml_index with
275 (log := !log ^ "Updating XML db ...<br />\n";
279 (match Pcre.split ~rex:index_line_sep_RE l with
281 assert (is_cic_uri uri || is_nuprl_uri uri) ;
282 (map_of_uri uri)#add uri ((xml_url_of_uri uri) ^ ".xml.gz")
284 assert (is_cic_uri uri || is_nuprl_uri uri) ;
285 (map_of_uri uri)#add uri ((xml_url_of_uri uri) ^ ".xml")
286 | _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "<br />\n")
287 with Http_getter_invalid_URI uri ->
288 log := !log ^ "Ignoring invalid XML URI: " ^ uri ^ "<br />\n")
289 (Pcre.split ~rex:index_sep_RE xml_index)) (* xml_index lines *)
291 (match rdf_index with
293 (log := !log ^ "Updating RDF db ...<br />\n";
297 (match Pcre.split ~rex:index_line_sep_RE l with
298 | [uri; "gz"] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml.gz")
299 | [uri] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml")
300 | _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "<br />\n")
301 with Http_getter_invalid_URI uri ->
302 log := !log ^ "Ignoring invalid RDF URI: " ^ uri ^ "<br />\n")
303 (Pcre.split ~rex:index_sep_RE rdf_index)) (* rdf_index lines *)
305 (match xsl_index with
307 (log := !log ^ "Updating XSLT db ...<br />\n";
309 (fun l -> xsl_map#add l (server_url ^ "/" ^ l))
310 (Pcre.split ~rex:index_sep_RE xsl_index);
311 log := !log ^ "All done!<br />\n")
313 debug_print "done with this server";
319 let callback (req: Http_types.request) outchan =
321 debug_print ("Connection from " ^ req#clientAddr);
322 debug_print ("Received request: " ^ req#path);
324 | "/help" -> return_html_raw Http_getter_const.usage_string outchan
325 | "/getxml" | "/getxslt" | "/getdtd" | "/resolve" | "/register" ->
326 (let uri = req#param "uri" in (* common parameter *)
329 let enc = parse_enc req in
331 try parse_patch req with Http_types.Param_not_found _ -> true
333 Http_getter_cache.respond_xml
334 ~url:(resolve uri) ~uri ~enc ~patch outchan
337 try parse_patch req with Http_types.Param_not_found _ -> true
339 Http_getter_cache.respond_xsl ~url:(resolve uri) ~patch outchan
342 try parse_patch req with Http_types.Param_not_found _ -> true
344 Http_getter_cache.respond_dtd
345 ~patch ~url:(Http_getter_env.dtd_dir ^ "/" ^ uri) outchan
349 (sprintf "<url value=\"%s\" />\n" (resolve uri))
351 with Http_getter_unresolvable_URI uri ->
352 return_xml_raw "<unresolved />\n" outchan)
354 let url = req#param "url" in
356 return_html_msg "Register done" outchan
359 (Http_getter_env.reload (); (* reload servers list from servers file *)
360 cic_map#clear; nuprl_map#clear; rdf_map#clear; xsl_map#clear;
364 "" (* initial logmsg: empty *)
365 (* reverse order: 1st server is the most important one *)
366 (List.rev !Http_getter_env.servers)
368 cic_map#sync; nuprl_map#sync; rdf_map#sync; xsl_map#sync;
369 return_html_msg log outchan)
373 (Pcre.pmatch ~rex:heading_cic_RE uri) &&
374 not (Pcre.pmatch ~rex:trailing_types_RE uri))
376 | "/getallrdfuris" ->
377 (let classs = req#param "class" in
380 let base = "^helm:rdf:www\\.cs\\.unibo\\.it/helm/rdf/" in
382 | ("forward" as c) | ("backward" as c) ->
383 (fun uri -> Pcre.pmatch ~pat:(base ^ c) uri)
384 | c -> raise (Http_getter_invalid_RDF_class c)
386 return_all_rdf_uris filter outchan
387 with Http_getter_invalid_RDF_class c ->
388 raise (Http_getter_bad_request ("Invalid RDF class: " ^ c)))
389 | "/ls" -> return_ls (parse_ls_uri req) (parse_output_format req) outchan
391 Http_daemon.respond ~body:Http_getter_const.empty_xml outchan
393 Http_daemon.respond_error ~status:(`Client_error `Bad_request) outchan);
394 debug_print "Done!\n"
396 | Http_types.Param_not_found attr_name ->
397 return_400 (sprintf "Parameter '%s' is missing" attr_name) outchan
398 | Http_getter_bad_request msg -> return_html_error msg outchan
399 | Http_getter_internal_error msg -> return_html_internal_error msg outchan
400 | Shell.Subprocess_error l ->
401 return_html_internal_error
402 (String.concat "<br />\n"
405 sprintf "Command '%s' returned %s"
406 cmd (string_of_proc_status code))
411 ("Uncaught exception: " ^ (Printexc.to_string exc))
415 (* daemon initialization *)
418 Http_getter_env.dump_env ();
419 Unix.putenv "http_proxy" "";
421 Sys.catch_break true;
424 ~timeout:(Some 600) ~port:Http_getter_env.port ~mode:`Thread callback
425 with Sys.Break -> () (* 'save_maps' already registered with 'at_exit' *)