X-Git-Url: http://matita.cs.unibo.it/gitweb/?a=blobdiff_plain;f=helm%2Fhttp_getter%2Fhttp_getter.ml;h=366c233b4979804322f4a83a1d03613a91312dfa;hb=4faf0e37e7019de16dd6862bb34d84f799a2a230;hp=02877b012c55690222fb7c2a87763225706f82bd;hpb=21a69c0525553c5e197ca0499705afe904fd7260;p=helm.git diff --git a/helm/http_getter/http_getter.ml b/helm/http_getter/http_getter.ml index 02877b012..366c233b4 100644 --- a/helm/http_getter/http_getter.ml +++ b/helm/http_getter/http_getter.ml @@ -1,5 +1,7 @@ (* - * Copyright (C) 2003, HELM Team. + * Copyright (C) 2003: + * Stefano Zacchiroli + * for the HELM Team http://helm.cs.unibo.it/ * * This file is part of HELM, an Hypertextual, Electronic * Library of Mathematics, developed at the Computer Science @@ -21,7 +23,7 @@ * MA 02111-1307, USA. * * For details, see the HELM World-Wide-Web page, - * http://cs.unibo.it/helm/. + * http://helm.cs.unibo.it/ *) open Http_getter_common;; @@ -40,6 +42,7 @@ let common_headers = [ (* HTTP queries argument parsing *) + (* parse encoding ("format" parameter), default is Enc_normal *) let parse_enc (req: Http_types.request) = try (match req#param "format" with @@ -48,42 +51,55 @@ let parse_enc (req: Http_types.request) = | s -> raise (Http_getter_bad_request ("Invalid format: " ^ s))) with Http_types.Param_not_found _ -> Enc_normal ;; -let parse_patch_dtd (req: Http_types.request) = - match req#param "patch_dtd" with - | s when String.lowercase s = "yes" -> true - | s when String.lowercase s = "no" -> false - | s -> raise (Http_getter_bad_request ("Invalid patch_dtd value: " ^ s)) + (* parse "patch_dtd" parameter, default is true *) +let parse_patch (req: Http_types.request) = + try + (match req#param "patch_dtd" with + | s when String.lowercase s = "yes" -> true + | s when String.lowercase s = "no" -> false + | s -> raise (Http_getter_bad_request ("Invalid patch_dtd value: " ^ s))) + with Http_types.Param_not_found _ -> true ;; + (* parse output format ("format" parameter), no default value *) let parse_output_format (req: Http_types.request) = match req#param "format" with | s when String.lowercase s = "txt" -> Fmt_text | s when String.lowercase s = "xml" -> Fmt_xml | s -> raise (Http_getter_bad_request ("Invalid /ls format: " ^ s)) ;; + (* parse "baseuri" format for /ls method, no default value *) let parse_ls_uri = let parse_ls_RE = Pcre.regexp "^(\\w+):(.*)$" in let trailing_slash_RE = Pcre.regexp "/+$" in + let wrong_uri uri = + raise (Http_getter_bad_request ("Invalid /ls baseuri: " ^ uri)) + in fun (req: Http_types.request) -> let baseuri = req#param "baseuri" in - let subs = - Pcre.extract ~rex:parse_ls_RE - (Pcre.replace ~rex:trailing_slash_RE baseuri) - in - match (subs.(1), subs.(2)) with - | "cic", uri -> Cic uri - | "theory", uri -> Theory uri - | _ -> raise (Http_getter_bad_request ("Invalid /ls baseuri: " ^ baseuri)) + try + let subs = + Pcre.extract ~rex:parse_ls_RE + (Pcre.replace ~rex:trailing_slash_RE baseuri) + in + (match (subs.(1), subs.(2)) with + | "cic", uri -> Cic uri + | "theory", uri -> Theory uri + | _ -> wrong_uri baseuri) + with Not_found -> wrong_uri baseuri ;; (* global maps, shared by all threads *) -let xml_map = new Http_getter_map.map Http_getter_env.xml_dbm in +let cic_map = new Http_getter_map.map Http_getter_env.cic_dbm in +let nuprl_map = new Http_getter_map.map Http_getter_env.nuprl_dbm in let rdf_map = new Http_getter_map.map Http_getter_env.rdf_dbm in let xsl_map = new Http_getter_map.map Http_getter_env.xsl_dbm in -let save_maps () = xml_map#close; rdf_map#close; xsl_map#close in +let save_maps () = + cic_map#close; nuprl_map#close; rdf_map#close; xsl_map#close in let map_of_uri = function - | uri when is_xml_uri uri -> xml_map + | uri when is_cic_uri uri -> cic_map + | uri when is_nuprl_uri uri -> nuprl_map | uri when is_rdf_uri uri -> rdf_map | uri when is_xsl_uri uri -> xsl_map | uri -> raise (Http_getter_unresolvable_URI uri) @@ -94,7 +110,10 @@ let resolve uri = with Http_getter_map.Key_not_found _ -> raise (Http_getter_unresolvable_URI uri) in -let register uri = (map_of_uri uri )#add uri in +let register uri = + (* Warning: this fail if uri is already registered *) + (map_of_uri uri)#add uri +in let return_all_foo_uris map doctype filter outchan = (** return all URIs contained in 'map' which satisfy predicate 'filter'; URIs are written in an XMLish format ('doctype' is the XML doctype) onto 'outchan' @@ -121,20 +140,25 @@ let return_all_foo_uris map doctype filter outchan = output_string outchan (sprintf "\t\n" uri)); output_string outchan (sprintf "\n" doctype) in -let return_all_xml_uris = return_all_foo_uris xml_map "alluris" in +let return_all_xml_uris = return_all_foo_uris cic_map "alluris" in let return_all_rdf_uris = return_all_foo_uris rdf_map "allrdfuris" in let return_ls = - let (++) (oldann, oldtypes, oldbody) (newann, newtypes, newbody) = + let (++) (oldann, oldtypes, oldbody, oldtree) + (newann, newtypes, newbody, newtree) = ((if newann > oldann then newann else oldann), (if newtypes > oldtypes then newtypes else oldtypes), - (if newbody > oldbody then newbody else oldbody)) - in + (if newbody > oldbody then newbody else oldbody), + (if newtree > oldtree then newtree else oldtree)) + in let basepart_RE = - Pcre.regexp "^([^.]*\\.[^.]*)((\\.body)|(\\.types))?(\\.ann)?" + Pcre.regexp + "^([^.]*\\.[^.]*)((\\.body)|(\\.proof_tree)|(\\.types))?(\\.ann)?$" in - let (types_RE, types_ann_RE, body_RE, body_ann_RE) = - (Pcre.regexp "\\.types", Pcre.regexp "\\.types.ann", - Pcre.regexp "\\.body", Pcre.regexp "\\.body.ann") + let (types_RE, types_ann_RE, body_RE, body_ann_RE, + proof_tree_RE, proof_tree_ann_RE) = + (Pcre.regexp "\\.types$", Pcre.regexp "\\.types\\.ann$", + Pcre.regexp "\\.body$", Pcre.regexp "\\.body\\.ann$", + Pcre.regexp "\\.proof_tree$", Pcre.regexp "\\.proof_tree\\.ann$") in let (slash_RE, til_slash_RE, no_slashes_RE) = (Pcre.regexp "/", Pcre.regexp "^.*/", Pcre.regexp "^[^/]*$") @@ -154,22 +178,26 @@ let return_ls = in let store_obj o = let basepart = Pcre.replace ~rex:basepart_RE ~templ:"$1" o in + let no_flags = false, No, No, No in let oldflags = try Hashtbl.find objs basepart - with Not_found -> (false, No, No) (* no ann, no types no body *) + with Not_found -> (* no ann, no types, no body, no proof tree *) + no_flags in let newflags = match o with - | s when Pcre.pmatch ~rex:types_RE s -> (false, Yes, No) - | s when Pcre.pmatch ~rex:types_ann_RE s -> (true, Ann, No) - | s when Pcre.pmatch ~rex:body_RE s -> (false, No, Yes) - | s when Pcre.pmatch ~rex:body_ann_RE s -> (true, No, Ann) - | s -> (false, No, No) + | s when Pcre.pmatch ~rex:types_RE s -> (false, Yes, No, No) + | s when Pcre.pmatch ~rex:types_ann_RE s -> (true, Ann, No, No) + | s when Pcre.pmatch ~rex:body_RE s -> (false, No, Yes, No) + | s when Pcre.pmatch ~rex:body_ann_RE s -> (true, No, Ann, No) + | s when Pcre.pmatch ~rex:proof_tree_RE s -> (false, No, No, Yes) + | s when Pcre.pmatch ~rex:proof_tree_ann_RE s -> (true, No, No, Ann) + | s -> no_flags in Hashtbl.replace objs basepart (oldflags ++ newflags) in - xml_map#iter (* BLEARGH Dbm module lacks support for fold-like functions *) + cic_map#iter (* BLEARGH Dbm module lacks support for fold-like functions *) (fun key _ -> match key with | uri when Pcre.pmatch ~rex:dir_RE uri -> (* directory hit *) @@ -188,15 +216,17 @@ let return_ls = (fun s d -> sprintf "%sdir, %s\n" s d) "" (StringSet.elements !dirs)) ^ (Http_getter_misc.hashtbl_sorted_fold - (fun uri (annflag, typesflag, bodyflag) cont -> - sprintf "%sobject, %s, <%s,%s,%s>\n" + (fun uri (annflag, typesflag, bodyflag, treeflag) cont -> + sprintf "%sobject, %s, <%s,%s,%s,%s>\n" cont uri (if annflag then "YES" else "NO") - (string_of_ls_flag typesflag) (string_of_ls_flag bodyflag)) + (string_of_ls_flag typesflag) + (string_of_ls_flag bodyflag) + (string_of_ls_flag treeflag)) objs "") in Http_daemon.respond ~headers:(("Content-Type", "text/plain") :: common_headers) - ~body outchan + ~body outchan | Fmt_xml -> let body = sprintf @@ -215,28 +245,31 @@ let return_ls = (fun d -> "
" ^ d ^ "
") (StringSet.elements !dirs))) ^ "\n" ^ (Http_getter_misc.hashtbl_sorted_fold - (fun uri (annflag, typesflag, bodyflag) cont -> + (fun uri (annflag, typesflag, bodyflag, treeflag) cont -> sprintf "%s \t \t \t +\t " cont uri (if annflag then "YES" else "NO") (string_of_ls_flag typesflag) - (string_of_ls_flag bodyflag)) + (string_of_ls_flag bodyflag) + (string_of_ls_flag treeflag)) objs "")) in Http_daemon.respond ~headers:(("Content-Type", "text/xml") :: common_headers) - ~body outchan + ~body outchan in let (index_line_sep_RE, index_sep_RE, trailing_types_RE, - heading_cic_RE, heading_theory_RE, + heading_cic_RE, heading_theory_RE, heading_nuprl_RE, heading_rdf_cic_RE, heading_rdf_theory_RE) = - (Pcre.regexp "[ \t]+", Pcre.regexp "\n+", Pcre.regexp "\\.types$", - Pcre.regexp "^cic:", Pcre.regexp "^theory:", + (Pcre.regexp "[ \t]+", Pcre.regexp "\r\n|\r|\n", + Pcre.regexp "\\.types$", + Pcre.regexp "^cic:", Pcre.regexp "^theory:", Pcre.regexp "^nuprl:", Pcre.regexp "^helm:rdf.*//cic:", Pcre.regexp "^helm:rdf.*//theory:") in let update_from_server logmsg server_url = (* use global maps *) @@ -247,6 +280,8 @@ let update_from_server logmsg server_url = (* use global maps *) Pcre.replace ~rex:heading_cic_RE ~templ:server_url uri | uri when (Pcre.pmatch ~rex:heading_theory_RE uri) -> Pcre.replace ~rex:heading_theory_RE ~templ:server_url uri + | uri when (Pcre.pmatch ~rex:heading_nuprl_RE uri) -> + Pcre.replace ~rex:heading_nuprl_RE ~templ:server_url uri | uri -> raise (Http_getter_invalid_URI uri) in let rdf_url_of_uri = function (* TODO as above *) @@ -269,15 +304,24 @@ let update_from_server logmsg server_url = (* use global maps *) | Some xml_index -> (log := !log ^ "Updating XML db ...
\n"; List.iter - (fun l -> - try - (match Pcre.split ~rex:index_line_sep_RE l with - | [uri; "gz"] -> xml_map#add uri ((xml_url_of_uri uri) ^ ".xml.gz") - | [uri] -> xml_map#add uri ((xml_url_of_uri uri) ^ ".xml") - | _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "
\n") - with Http_getter_invalid_URI uri -> - log := !log ^ "Ignoring invalid XML URI: " ^ uri ^ "
\n") - (Pcre.split ~rex:index_sep_RE xml_index)) (* xml_index lines *) + (function + | l when is_blank_line l -> () (* skip blank and commented lines *) + | l -> + try + (match Pcre.split ~rex:index_line_sep_RE l with + | [uri; "gz"] -> + assert (is_cic_uri uri || is_nuprl_uri uri) ; + (map_of_uri uri)#replace + uri ((xml_url_of_uri uri) ^ ".xml.gz") + | [uri] -> + assert (is_cic_uri uri || is_nuprl_uri uri) ; + (map_of_uri uri)#replace + uri ((xml_url_of_uri uri) ^ ".xml") + | _ -> + log := !log ^ "Ignoring invalid line: '" ^ l ^ "'
\n") + with Http_getter_invalid_URI uri -> + log := !log ^ "Ignoring invalid XML URI: '" ^ uri ^ "'
\n") + (Pcre.split ~rex:index_sep_RE xml_index)) (* xml_index lines *) | None -> ()); (match rdf_index with | Some rdf_index -> @@ -286,8 +330,9 @@ let update_from_server logmsg server_url = (* use global maps *) (fun l -> try (match Pcre.split ~rex:index_line_sep_RE l with - | [uri; "gz"] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml.gz") - | [uri] -> rdf_map#add uri ((rdf_url_of_uri uri) ^ ".xml") + | [uri; "gz"] -> + rdf_map#replace uri ((rdf_url_of_uri uri) ^ ".xml.gz") + | [uri] -> rdf_map#replace uri ((rdf_url_of_uri uri) ^ ".xml") | _ -> log := !log ^ "Ignoring invalid line: " ^ l ^ "
\n") with Http_getter_invalid_URI uri -> log := !log ^ "Ignoring invalid RDF URI: " ^ uri ^ "
\n") @@ -297,13 +342,25 @@ let update_from_server logmsg server_url = (* use global maps *) | Some xsl_index -> (log := !log ^ "Updating XSLT db ...
\n"; List.iter - (fun l -> xsl_map#add l (server_url ^ "/" ^ l)) + (fun l -> xsl_map#replace l (server_url ^ "/" ^ l)) (Pcre.split ~rex:index_sep_RE xsl_index); log := !log ^ "All done!
\n") | None -> ()); debug_print "done with this server"; !log in +let update_from_all_servers () = (* use global maps *) + cic_map#clear; nuprl_map#clear; rdf_map#clear; xsl_map#clear; + let log = + List.fold_left + update_from_server + "" (* initial logmsg: empty *) + (* reverse order: 1st server is the most important one *) + (List.rev !Http_getter_env.servers) + in + cic_map#sync; nuprl_map#sync; rdf_map#sync; xsl_map#sync; + log +in (* thread action *) @@ -312,51 +369,109 @@ let callback (req: Http_types.request) outchan = debug_print ("Connection from " ^ req#clientAddr); debug_print ("Received request: " ^ req#path); (match req#path with - | "/help" -> return_html_msg Http_getter_const.usage_string outchan + | "/help" -> + return_html_raw + (Http_getter_const.usage_string (Http_getter_env.env_to_string ())) + outchan | "/getxml" | "/getxslt" | "/getdtd" | "/resolve" | "/register" -> (let uri = req#param "uri" in (* common parameter *) match req#path with | "/getxml" -> let enc = parse_enc req in - let patch_dtd = - try parse_patch_dtd req with Http_types.Param_not_found _ -> true - in + let patch = parse_patch req in Http_getter_cache.respond_xml - ~url:(resolve uri) ~uri ~enc ~patch_dtd outchan + ~url:(resolve uri) ~uri ~enc ~patch outchan | "/getxslt" -> - let patch_dtd = - try parse_patch_dtd req with Http_types.Param_not_found _ -> true - in - Http_getter_cache.respond_xsl ~url:(resolve uri) ~patch_dtd outchan + let patch = parse_patch req in + Http_getter_cache.respond_xsl ~url:(resolve uri) ~patch outchan | "/getdtd" -> - let patch_dtd = - try parse_patch_dtd req with Http_types.Param_not_found _ -> true - in + let patch = parse_patch req in Http_getter_cache.respond_dtd - ~patch_dtd ~url:(Http_getter_env.dtd_dir ^ "/" ^ uri) outchan + ~patch ~url:(Http_getter_env.dtd_dir ^ "/" ^ uri) outchan | "/resolve" -> (try - return_xml_msg + return_xml_raw (sprintf "\n" (resolve uri)) outchan with Http_getter_unresolvable_URI uri -> - return_xml_msg "\n" outchan) + return_xml_raw "\n" outchan) | "/register" -> let url = req#param "url" in register uri url; return_html_msg "Register done" outchan | _ -> assert false) + | "/clean_cache" -> + Http_getter_cache.clean (); + return_html_msg "Done." outchan | "/update" -> - (xml_map#clear; rdf_map#clear; xsl_map#clear; - let log = - List.fold_left - update_from_server - "" (* initial logmsg: empty *) - (* reverse order: 1st server is the most important one *) - (List.rev Http_getter_env.servers) + Http_getter_env.reload (); (* reload servers list from servers file *) + let log = update_from_all_servers () in + return_html_msg log outchan + | "/list_servers" -> + return_html_raw + (sprintf "\n%s\n
" + (String.concat "\n" + (List.map + (let i = ref ~-1 in + fun s -> incr i; sprintf "%d%s" !i s) + !Http_getter_env.servers))) + outchan + | "/add_server" -> + let name = req#param "url" in + (try + let position = + try + let res = int_of_string (req#param "position") in + if res < 0 then + raise (Failure "int_of_string"); + res + with Failure "int_of_string" -> + raise (Http_getter_bad_request + (sprintf "position must be a non negative integer (%s given)" + (req#param "position"))) + in + if position = 0 then (* fallback to default value *) + raise (Http_types.Param_not_found "foo") + else if position > 0 then begin (* add server and update all *) + Http_getter_env.add_server ~position name; + let log = update_from_all_servers () in + return_html_msg + (sprintf "Added server %s in position %d)
\n%s" + name position log) + outchan + end else (* position < 0 *) (* error! *) + assert false (* already checked above *) + with Http_types.Param_not_found _ -> (* add as 1st server by default *) + Http_getter_env.add_server ~position:0 name; + let log = update_from_server (* quick update (new server only) *) + (sprintf "Added server %s in head position
\n" name) name + in + return_html_msg log outchan) + | "/remove_server" -> + let position = + try + let res = int_of_string (req#param "position") in + if res < 0 then + raise (Failure "int_of_string"); + res + with Failure "int_of_string" -> + raise (Http_getter_bad_request + (sprintf "position must be a non negative integer (%s given)" + (req#param "position"))) in - xml_map#sync; rdf_map#sync; xsl_map#sync; - return_html_msg log outchan) + let server_name = + try + List.nth !Http_getter_env.servers position + with Failure "nth" -> + raise (Http_getter_bad_request + (sprintf "no server with position %d" position)) + in + Http_getter_env.remove_server position; + let log = update_from_all_servers () in + return_html_msg + (sprintf "Removed server %s (position %d)
\n%s" + server_name position log) + outchan | "/getalluris" -> return_all_xml_uris (fun uri -> @@ -405,7 +520,8 @@ in (* daemon initialization *) let main () = - Http_getter_env.dump_env (); + print_string (Http_getter_env.env_to_string ()); + flush stdout; Unix.putenv "http_proxy" ""; at_exit save_maps; Sys.catch_break true;