2 * Copyright (C) 2003, HELM Team.
4 * This file is part of HELM, an Hypertextual, Electronic
5 * Library of Mathematics, developed at the Computer Science
6 * Department, University of Bologna, Italy.
8 * HELM is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU General Public License
10 * as published by the Free Software Foundation; either version 2
11 * of the License, or (at your option) any later version.
13 * HELM is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
18 * You should have received a copy of the GNU General Public License
19 * along with HELM; if not, write to the Free Software
20 * Foundation, Inc., 59 Temple Place - Suite 330, Boston,
23 * For details, see the HELM World-Wide-Web page,
24 * http://cs.unibo.it/helm/.
27 open Http_getter_common;;
28 open Http_getter_misc;;
29 open Http_getter_types;;
30 open Http_getter_debugger;;
35 let common_headers = [
36 "Cache-Control", "no-cache";
41 (* HTTP queries argument parsing *)
43 let parse_enc (req: Http_types.request) =
45 (match req#param "format" with
46 | "normal" -> Enc_normal
48 | s -> raise (Http_getter_bad_request ("Invalid format: " ^ s)))
49 with Http_types.Param_not_found _ -> Enc_normal
51 let parse_patch_dtd (req: Http_types.request) =
52 match req#param "patch_dtd" with
53 | s when String.lowercase s = "yes" -> true
54 | s when String.lowercase s = "no" -> false
55 | s -> raise (Http_getter_bad_request ("Invalid patch_dtd value: " ^ s))
57 let parse_output_format (req: Http_types.request) =
58 match req#param "format" with
59 | s when String.lowercase s = "txt" -> Fmt_text
60 | s when String.lowercase s = "xml" -> Fmt_xml
61 | s -> raise (Http_getter_bad_request ("Invalid /ls format: " ^ s))
64 let parse_ls_RE = Pcre.regexp "^(\\w+):(.*)$" in
65 let trailing_slash_RE = Pcre.regexp "/+$" in
66 fun (req: Http_types.request) ->
67 let baseuri = req#param "baseuri" in
69 Pcre.extract ~rex:parse_ls_RE
70 (Pcre.replace ~rex:trailing_slash_RE baseuri)
72 match (subs.(1), subs.(2)) with
73 | "cic", uri -> Cic uri
74 | "theory", uri -> Theory uri
75 | _ -> raise (Http_getter_bad_request ("Invalid /ls baseuri: " ^ baseuri))
78 (* global maps, shared by all threads *)
80 let xml_map = new Http_getter_map.map Http_getter_env.xml_dbm in
81 let rdf_map = new Http_getter_map.map Http_getter_env.rdf_dbm in
82 let xsl_map = new Http_getter_map.map Http_getter_env.xsl_dbm in
84 let save_maps () = xml_map#close; rdf_map#close; xsl_map#close in
85 let map_of_uri = function
86 | uri when is_xml_uri uri -> xml_map
87 | uri when is_rdf_uri uri -> rdf_map
88 | uri when is_xsl_uri uri -> xsl_map
89 | uri -> raise (Http_getter_unresolvable_URI uri)
91 let resolve uri = (map_of_uri uri)#resolve uri in
92 let register uri = (map_of_uri uri )#add uri in
93 let return_all_foo_uris map doctype filter outchan =
94 (** return all URIs contained in 'map' which satisfy predicate 'filter'; URIs
95 are written in an XMLish format ('doctype' is the XML doctype) onto 'outchan'
97 Http_daemon.send_basic_headers ~code:200 outchan;
98 Http_daemon.send_header "Content-Type" "text/xml" outchan;
99 Http_daemon.send_headers common_headers outchan;
100 Http_daemon.send_CRLF outchan;
104 "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>
105 <!DOCTYPE %s SYSTEM \"%s/getdtd?uri=%s.dtd\">
110 Http_getter_env.my_own_url
116 output_string outchan (sprintf "\t<uri value=\"%s\" />\n" uri));
117 output_string outchan (sprintf "</%s>\n" doctype)
119 let return_all_xml_uris = return_all_foo_uris xml_map "alluris" in
120 let return_all_rdf_uris = return_all_foo_uris rdf_map "allrdfuris" in
122 let (++) (oldann, oldtypes, oldbody) (newann, newtypes, newbody) =
123 ((if newann > oldann then newann else oldann),
124 (if newtypes > oldtypes then newtypes else oldtypes),
125 (if newbody > oldbody then newbody else oldbody))
128 Pcre.regexp "^([^.]*\\.[^.]*)((\\.body)|(\\.types))?(\\.ann)?"
130 let (types_RE, types_ann_RE, body_RE, body_ann_RE) =
131 (Pcre.regexp "\\.types", Pcre.regexp "\\.types.ann",
132 Pcre.regexp "\\.body", Pcre.regexp "\\.body.ann")
134 let (slash_RE, til_slash_RE, no_slashes_RE) =
135 (Pcre.regexp "/", Pcre.regexp "^.*/", Pcre.regexp "^[^/]*$")
137 fun lsuri fmt outchan ->
140 (match lsuri with Cic p -> ("cic:" ^ p) | Theory p -> ("theory:" ^ p))
142 let (dir_RE, obj_RE) =
143 (Pcre.regexp (pat ^ "/"), Pcre.regexp (pat ^ "(\\.|$)"))
145 let dirs = ref StringSet.empty in
146 let objs = Hashtbl.create 17 in
148 dirs := StringSet.add (List.hd (Pcre.split ~rex:slash_RE d)) !dirs
151 let basepart = Pcre.replace ~rex:basepart_RE ~templ:"$1" o in
154 Hashtbl.find objs basepart
155 with Not_found -> (false, No, No) (* no ann, no types no body *)
159 | s when Pcre.pmatch ~rex:types_RE s -> (false, Yes, No)
160 | s when Pcre.pmatch ~rex:types_ann_RE s -> (true, Ann, No)
161 | s when Pcre.pmatch ~rex:body_RE s -> (false, No, Yes)
162 | s when Pcre.pmatch ~rex:body_ann_RE s -> (true, No, Ann)
163 | s -> (false, No, No)
165 Hashtbl.replace objs basepart (oldflags ++ newflags)
167 xml_map#iter (* BLEARGH Dbm module lacks support for fold-like functions *)
170 | uri when Pcre.pmatch ~rex:dir_RE uri -> (* directory hit *)
171 let localpart = Pcre.replace ~rex:dir_RE uri in
172 if Pcre.pmatch ~rex:no_slashes_RE localpart then
176 | uri when Pcre.pmatch ~rex:obj_RE uri -> (* file hit *)
177 store_obj (Pcre.replace ~rex:til_slash_RE uri)
178 | uri -> () (* miss *));
183 (fun s d -> sprintf "%sdir, %s\n" s d) ""
184 (StringSet.elements !dirs)) ^
185 (Http_getter_misc.hashtbl_sorted_fold
186 (fun uri (annflag, typesflag, bodyflag) cont ->
187 sprintf "%sobject, %s, <%s,%s,%s>\n"
188 cont uri (if annflag then "YES" else "NO")
189 (string_of_ls_flag typesflag) (string_of_ls_flag bodyflag))
193 ~headers:(("Content-Type", "text/plain") :: common_headers)
198 "<?xml version=\"1.0\" encoding=\"ISO-8859-1\"?>
199 <!DOCTYPE ls SYSTEM \"%s/getdtd?uri=ls.dtd\">
205 Http_getter_env.my_own_url
210 (fun d -> "<section>" ^ d ^ "</section>")
211 (StringSet.elements !dirs))) ^ "\n" ^
212 (Http_getter_misc.hashtbl_sorted_fold
213 (fun uri (annflag, typesflag, bodyflag) cont ->
215 "%s<object name=\"%s\">
216 \t<ann value=\"%s\" />
217 \t<types value=\"%s\" />
218 \t<body value=\"%s\" />
221 cont uri (if annflag then "YES" else "NO")
222 (string_of_ls_flag typesflag)
223 (string_of_ls_flag bodyflag))
227 ~headers:(("Content-Type", "text/xml") :: common_headers)
230 let (index_line_sep_RE, index_sep_RE, trailing_types_RE,
231 heading_cic_RE, heading_theory_RE,
232 heading_rdf_cic_RE, heading_rdf_theory_RE) =
233 (Pcre.regexp "[ \t]+", Pcre.regexp "\n+", Pcre.regexp "\\.types$",
234 Pcre.regexp "^cic:", Pcre.regexp "^theory:",
235 Pcre.regexp "^helm:rdf.*//cic:", Pcre.regexp "^helm:rdf.*//theory:")
237 let update_from_server logmsg server_url = (* use global maps *)
238 debug_print ("Updating information from " ^ server_url);
239 let xml_url_of_uri = function
240 (* TODO missing sanity checks on server_url, e.g. it can contains $1 *)
241 | uri when (Pcre.pmatch ~rex:heading_cic_RE uri) ->
242 Pcre.replace ~rex:heading_cic_RE ~templ:server_url uri
243 | uri when (Pcre.pmatch ~rex:heading_theory_RE uri) ->
244 Pcre.replace ~rex:heading_theory_RE ~templ:server_url uri
245 | uri -> raise (Http_getter_invalid_URI uri)
247 let rdf_url_of_uri = function (* TODO as above *)
248 | uri when (Pcre.pmatch ~rex:heading_rdf_cic_RE uri) ->
249 Pcre.replace ~rex:heading_rdf_cic_RE ~templ:server_url uri
250 | uri when (Pcre.pmatch ~rex:heading_rdf_theory_RE uri) ->
251 Pcre.replace ~rex:heading_rdf_theory_RE ~templ:server_url uri
252 | uri -> raise (Http_getter_invalid_URI uri)
254 let log = ref (logmsg ^ "Processing server: " ^ server_url ^ "<br />\n") in
255 let (xml_index, rdf_index, xsl_index) =
256 (* TODO keeps index in memory, is better to keep them on temp files? *)
257 (http_get (server_url ^ "/" ^ Http_getter_env.xml_index),
258 http_get (server_url ^ "/" ^ Http_getter_env.rdf_index),
259 http_get (server_url ^ "/" ^ Http_getter_env.xsl_index))
261 if (xml_index = None && rdf_index = None && xsl_index = None) then
262 debug_print (sprintf "Warning: useless server %s" server_url);
263 (match xml_index with
265 (log := !log ^ "Updating XML db ...<br />\n";
269 (match Pcre.split ~rex:index_line_sep_RE l with
270 | [uri; "gz"] -> xml_map#add uri ((xml_url_of_uri uri) ^ ".xml.gz")
271 | [uri] -> xml_map#add uri ((xml_url_of_uri uri) ^ ".xml")
272 | _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "<br />\n")
273 with Http_getter_invalid_URI uri ->
274 log := !log ^ "Ignoring invalid XML URI: " ^ uri ^ "<br />\n")
275 (Pcre.split ~rex:index_sep_RE xml_index)) (* xml_index lines *)
277 (match rdf_index with
279 (log := !log ^ "Updating RDF db ...<br />\n";
283 (match Pcre.split ~rex:index_line_sep_RE l with
284 | [uri; "gz"] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml.gz")
285 | [uri] -> rdf_map#add uri ((rdf_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 RDF URI: " ^ uri ^ "<br />\n")
289 (Pcre.split ~rex:index_sep_RE rdf_index)) (* rdf_index lines *)
291 (match xsl_index with
293 (log := !log ^ "Updating XSLT db ...<br />\n";
295 (fun l -> xsl_map#add l (server_url ^ "/" ^ l))
296 (Pcre.split ~rex:index_sep_RE xsl_index);
297 log := !log ^ "All done!<br />\n")
299 debug_print "done with this server";
305 let callback (req: Http_types.request) outchan =
307 debug_print ("Connection from " ^ req#clientAddr);
308 debug_print ("Received request: " ^ req#path);
310 | "/help" -> return_html_msg Http_getter_const.usage_string outchan
311 | "/getxml" | "/getxslt" | "/getdtd" | "/resolve" | "/register" ->
312 (let uri = req#param "uri" in (* common parameter *)
315 let enc = parse_enc req in
317 try parse_patch_dtd req with Http_types.Param_not_found _ -> true
319 Http_getter_cache.respond_xml
320 ~url:(resolve uri) ~uri ~enc ~patch_dtd outchan
323 try parse_patch_dtd req with Http_types.Param_not_found _ -> true
325 Http_getter_cache.respond_xsl ~url:(resolve uri) ~patch_dtd outchan
328 try parse_patch_dtd req with Http_types.Param_not_found _ -> true
330 Http_getter_cache.respond_dtd
331 ~patch_dtd ~url:(Http_getter_env.dtd_dir ^ "/" ^ uri) outchan
335 (sprintf "<url value=\"%s\" />\n" (resolve uri))
337 with Http_getter_unresolvable_URI uri ->
338 return_xml_msg "<unresolved />\n" outchan)
340 let url = req#param "url" in
342 return_html_msg "Register done" outchan
345 (xml_map#clear; rdf_map#clear; xsl_map#clear;
349 "" (* initial logmsg: empty *)
350 (* reverse order: 1st server is the most important one *)
351 (List.rev Http_getter_env.servers)
353 xml_map#sync; rdf_map#sync; xsl_map#sync;
354 return_html_msg log outchan)
358 (Pcre.pmatch ~rex:heading_cic_RE uri) &&
359 not (Pcre.pmatch ~rex:trailing_types_RE uri))
361 | "/getallrdfuris" ->
362 (let classs = req#param "class" in
365 let base = "^helm:rdf:www\\.cs\\.unibo\\.it/helm/rdf/" in
367 | ("forward" as c) | ("backward" as c) ->
368 (fun uri -> Pcre.pmatch ~pat:(base ^ c) uri)
369 | c -> raise (Http_getter_invalid_RDF_class c)
371 return_all_rdf_uris filter outchan
372 with Http_getter_invalid_RDF_class c ->
373 raise (Http_getter_bad_request ("Invalid RDF class: " ^ c)))
374 | "/ls" -> return_ls (parse_ls_uri req) (parse_output_format req) outchan
376 Http_daemon.respond ~body:Http_getter_const.empty_xml outchan
378 Http_daemon.respond_error ~status:(`Client_error `Bad_request) outchan);
379 debug_print "Done!\n"
381 | Http_types.Param_not_found attr_name ->
382 return_400 (sprintf "Parameter '%s' is missing" attr_name) outchan
383 | Http_getter_bad_request msg -> return_html_error msg outchan
384 | Http_getter_internal_error msg -> return_html_internal_error msg outchan
385 | Shell.Subprocess_error l ->
386 return_html_internal_error
387 (String.concat "<br />\n"
390 sprintf "Command '%s' returned %s"
391 cmd (string_of_proc_status code))
396 ("Uncaught exception: " ^ (Printexc.to_string exc))
400 (* daemon initialization *)
403 Http_getter_env.dump_env ();
404 Unix.putenv "http_proxy" "";
406 Sys.catch_break true;
409 ~timeout:(Some 600) ~port:Http_getter_env.port ~mode:`Thread callback
410 with Sys.Break -> () (* 'save_maps' already registered with 'at_exit' *)