]> matita.cs.unibo.it Git - helm.git/blobdiff - helm/searchEngine/searchEngine.ml
- New interface for the MathQL interpreter (1.3 version)
[helm.git] / helm / searchEngine / searchEngine.ml
index 7b5635704f782e4b19cdfb4db942fafc3777b3f0..5e742b243a6b9d04c0b1f6b96a5c89d92e49dda4 100644 (file)
@@ -1,4 +1,3 @@
-
 (* Copyright (C) 2002, HELM Team.
  * 
  * This file is part of HELM, an Hypertextual, Electronic
  * http://cs.unibo.it/helm/.
  *)
 
+open Http_types ;;
+
 let debug = true;;
 let debug_print s = if debug then prerr_endline s;;
 Http_common.debug := true;;
+(* Http_common.debug := true;; *)
+
+  (** accepted HTTP servers for ask_uwobo method forwarding *)
+let valid_servers = [ "mowgli.cs.unibo.it:58080" ; "mowgli.cs.unibo.it" ; "localhost:58080" ] ;;
+
+module MQICallbacks =
+   struct
+      let log s = debug_print s
+   end
+
+module MQI = MQueryInterpreter.Make(MQICallbacks) 
+
+let mqi_options = "" (* default MathQL interpreter options *)
 
 open Printf;;
 
@@ -34,13 +48,39 @@ let daemon_name = "Search Engine";;
 let default_port = 58085;;
 let port_env_var = "SEARCH_ENGINE_PORT";;
 
-let pages_dir = "html";; (* relative to searchEngine's document root *)
+let pages_dir =
+  try
+    Sys.getenv "SEARCH_ENGINE_HTML_DIR"
+  with Not_found -> "html"  (* relative to searchEngine's document root *)
+;;
 let interactive_user_uri_choice_TPL = pages_dir ^ "/templateambigpdq1.html";;
 let interactive_interpretation_choice_TPL = pages_dir ^ "/templateambigpdq2.html";;
 let final_results_TPL = pages_dir ^ "/templateambigpdq3.html";;
 
 exception Chat_unfinished
 
+  (** pretty print a MathQL query result to an HELM theory file *)
+let theory_of_result result =
+ let results_no = List.length result in
+  if results_no > 0 then
+   let mode = if results_no > 10 then "linkonly" else "typeonly" in
+   let results =
+    let idx = ref (results_no + 1) in
+     List.fold_right
+      (fun (uri,attrs) i ->
+        decr idx ;
+        "<tr><td valign=\"top\">" ^ string_of_int !idx ^ ".</td><td><ht:OBJECT uri=\"" ^ uri ^ "\" mode=\"" ^ mode ^ "\"/></td></tr>" ^  i
+      ) result ""
+   in
+    "<h1>Query Results:</h1><table xmlns:ht=\"http://www.cs.unibo.it/helm/namespaces/helm-theory\">" ^ results ^ "</table>"
+  else
+    "<h1>Query Results:</h1><p>No results found!</p>"
+;;
+
+let pp_result result =
+ "<html xmlns:ht=\"http://www.cs.unibo.it/helm/namespaces/helm-theory\">\n<head><title>Query Results</title><style> A { text-decoration: none } </style></head>\n<body>" ^ theory_of_result result ^ "</body></html>"
+;;
+
   (** chain application of Pcre substitutions *)
 let rec apply_substs substs line =
   match substs with
@@ -62,11 +102,15 @@ let fold_file f init fname =
 let iter_file f = fold_file (fun _ line -> f line) ()
 
 let (title_tag_RE, choices_tag_RE, msg_tag_RE, id_to_uris_RE, id_RE,
-    interpretations_RE, interpretations_labels_RE, results_RE, new_aliases_RE) =
+    interpretations_RE, interpretations_labels_RE, results_RE, new_aliases_RE)
+  =
   (Pcre.regexp "@TITLE@", Pcre.regexp "@CHOICES@", Pcre.regexp "@MSG@",
   Pcre.regexp "@ID_TO_URIS@", Pcre.regexp "@ID@",
   Pcre.regexp "@INTERPRETATIONS@", Pcre.regexp "@INTERPRETATIONS_LABELS@",
   Pcre.regexp "@RESULTS@", Pcre.regexp "@NEW_ALIASES@")
+let server_and_port_url_RE = Pcre.regexp "^http://([^/]+)/.*$"
+
+exception NotAnInductiveDefinition
 
 let port =
   try
@@ -77,11 +121,6 @@ let port =
       prerr_endline "Warning: invalid port, reverting to default";
       default_port
 in
-let pp_result result =
-  let result_string = MQueryUtil.text_of_result result "\n" in
-  (sprintf "<html>\n<head>\n</head>\n<body>\n<pre>%s</pre>\n</body>\n</html>"
-    result_string)
-in
 let pp_error = sprintf "<html><body><h1>Error: %s</h1></body></html>" in
 let bad_request body outchan =
   Http_daemon.respond_error ~status:(`Client_error `Bad_request) ~body outchan
@@ -90,37 +129,129 @@ let contype = "Content-Type", "text/html" in
 
 (* SEARCH ENGINE functions *)
 
-let refine_constraints (x, y, z) = (x, y, z), (Some x, Some y, Some z) in
+let refine_constraints (constr_obj, constr_rel, constr_sort) =
+ function
+    "/searchPattern" ->
+      (constr_obj, constr_rel, constr_sort),
+       (Some constr_obj, Some constr_rel, Some constr_sort)
+  | "/matchConclusion" ->
+      let constr_obj' =
+       List.map
+        (function (uri,pos,_) -> (uri,pos,None))
+        (List.filter
+          (function (uri,pos,depth) as constr ->
+            pos="http://www.cs.unibo.it/helm/schemas/schema-helm#MainConclusion"
+            or
+            pos="http://www.cs.unibo.it/helm/schemas/schema-helm#InConclusion"
+          ) constr_obj)
+      in
+       (*CSC: we must select the must constraints here!!! *)
+       (constr_obj',[],[]),(Some constr_obj', None, None)
+  | _ -> assert false
+in
+
+let get_constraints term =
+ function
+    "/locateInductivePrinciple" ->
+      let uri = 
+       match term with
+          Cic.MutInd (uri,t,_) -> MQueryUtil.string_of_uriref (uri,[t])
+        | _ -> raise NotAnInductiveDefinition
+      in
+      let constr_obj =
+       [uri,"http://www.cs.unibo.it/helm/schemas/schema-helm#InHypothesis",
+         None ;
+        uri,"http://www.cs.unibo.it/helm/schemas/schema-helm#MainHypothesis",
+         Some 0
+       ]
+      in
+      let constr_rel =
+       ["http://www.cs.unibo.it/helm/schemas/schema-helm#MainConclusion",
+        None] in
+      let constr_sort =
+       ["http://www.cs.unibo.it/helm/schemas/schema-helm#MainHypothesis",
+        Some 1, "http://www.cs.unibo.it/helm/schemas/schema-helm#Prop"]
+      in
+       (constr_obj, constr_rel, constr_sort), (None,None,None)
+  | req_path ->
+     let must = MQueryLevels2.get_constraints term in
+      refine_constraints must req_path
+in
 
 (* HTTP DAEMON CALLBACK *)
 
 let callback (req: Http_types.request) outchan =
   try
     debug_print (sprintf "Received request: %s" req#path);
+    if req#path <> "/getpage" then
+      ignore (MQI.init mqi_options);
     (match req#path with
     | "/execute" ->
         let query_string = req#param "query" in
         let lexbuf = Lexing.from_string query_string in
         let query = MQueryUtil.query_of_text lexbuf in
         let result = MQueryGenerator.execute_query query in
-        let result_string = MQueryUtil.text_of_result result "\n" in
-        Http_daemon.respond
-          ~body:
-            (sprintf "<html><body><pre>%s</pre></body></html>" result_string)
-          ~headers:[contype] outchan
+        let result_string = pp_result result in
+        Http_daemon.respond ~body:result_string ~headers:[contype] outchan
     | "/locate" ->
         let id = req#param "id" in
         let result = MQueryGenerator.locate id in
         Http_daemon.respond ~headers:[contype] ~body:(pp_result result) outchan
     | "/getpage" ->
+        (* TODO implement "is_permitted" *)
         (let is_permitted _ = true in
         let remove_fragment uri = Pcre.replace ~pat:"#.*" uri in
-        match req#param "url" with
-        | page when is_permitted (remove_fragment page) ->
-            Http_daemon.respond_file
-              ~fname:(sprintf "%s/%s" pages_dir (remove_fragment page)) outchan
-        | page -> Http_daemon.respond_forbidden ~url:page outchan)
-    | "/searchPattern" ->
+        let page = remove_fragment (req#param "url") in
+        let preprocess =
+          (try
+            bool_of_string (req#param "preprocess")
+          with Invalid_argument _ | Http_types.Param_not_found _ -> false)
+        in
+        (match page with
+        | page when is_permitted page ->
+            (let fname = sprintf "%s/%s" pages_dir (remove_fragment page) in
+            Http_daemon.send_basic_headers ~code:200 outchan;
+            Http_daemon.send_header "Content-Type" "text/html" outchan;
+            Http_daemon.send_CRLF outchan;
+            if preprocess then begin
+              iter_file
+                (fun line ->
+                  output_string outchan
+                    ((apply_substs
+                       (List.map
+                         (function (key,value) ->
+                           let key' =
+                            (Pcre.extract ~pat:"param\\.(.*)" key).(1)
+                           in
+                            Pcre.regexp ("@" ^ key' ^ "@"), value
+                         )
+                         (List.filter
+                           (fun (key,_) as p-> Pcre.pmatch ~pat:"^param\\." key)
+                           req#params)
+                       )
+                       line) ^
+                    "\n"))
+                fname
+            end else
+              Http_daemon.send_file ~src:(FileSrc fname) outchan)
+        | page -> Http_daemon.respond_forbidden ~url:page outchan))
+    | "/ask_uwobo" ->
+      let url = req#param "url" in
+      let server_and_port =
+        (Pcre.extract ~rex:server_and_port_url_RE url).(1)
+      in
+      if List.mem server_and_port valid_servers then
+        Http_daemon.respond
+          ~headers:["Content-Type", "text/html"]
+          ~body:(Http_client.Convenience.http_get url)
+          outchan
+      else
+        Http_daemon.respond
+          ~body:(pp_error ("Untrusted UWOBO server: " ^ server_and_port))
+          outchan
+    | "/searchPattern"
+    | "/matchConclusion"
+    | "/locateInductivePrinciple" ->
         let term_string = req#param "term" in
         let lexbuf = Lexing.from_string term_string in
         let (context, metasenv) = ([], []) in
@@ -133,13 +264,15 @@ let callback (req: Http_types.request) outchan =
         let rec parse_tokens keys lookup = function (* TODO spostarla fuori *)
           | [] -> keys, lookup
           | "alias" :: key :: value :: rest ->
-              parse_tokens
-                (key::keys)
-                (fun id ->
-                  if id = key then
-                    Some (Disambiguate.cic_textual_parser_uri_of_string value)
-                  else lookup id)
-                rest
+              let key' = CicTextualParser0.Id key in
+               parse_tokens
+                 (key'::keys)
+                 (fun id ->
+                   if id = key' then
+                     Some
+                      (CicTextualParser0.Uri (MQueryMisc.cic_textual_parser_uri_of_string value))
+                   else lookup id)
+                 rest
           | _ -> failwith "Can't parse aliases"
         in
         let parse_choices choices_raw =
@@ -150,6 +283,8 @@ let callback (req: Http_types.request) outchan =
               | ""::id::tail
               | id::tail when id<>"" ->
                   (fun id' ->
+prerr_endline ("#### " ^ id ^ " :=");
+List.iter (fun u -> prerr_endline ("<" ^ Netencoding.Url.decode u ^ ">")) tail;
                     if id = id' then
                       Some (List.map (fun u -> Netencoding.Url.decode u) tail)
                     else
@@ -158,7 +293,8 @@ let callback (req: Http_types.request) outchan =
             (fun _ -> None)
             choices
         in
-        let id_to_uris = parse_tokens [] (fun _ -> None) tokens in
+        let (id_to_uris : Disambiguate.domain_and_interpretation) =
+         parse_tokens [] (fun _ -> None) tokens in
         let id_to_choices =
           try
             let choices_raw = req#param "choices" in
@@ -168,6 +304,12 @@ let callback (req: Http_types.request) outchan =
         let module Chat: Disambiguate.Callbacks =
           struct
 
+            let get_metasenv () =
+             !CicTextualParser0.metasenv
+
+            let set_metasenv metasenv =
+              CicTextualParser0.metasenv := metasenv
+
             let output_html = prerr_endline
 
             let interactive_user_uri_choice
@@ -182,6 +324,8 @@ let callback (req: Http_types.request) outchan =
                   (match selection_mode with
                   | `SINGLE -> assert false
                   | `EXTENDED ->
+                      Http_daemon.send_basic_headers ~code:200 outchan ;
+                      Http_daemon.send_CRLF outchan ;
                       iter_file
                         (fun line ->
                           let formatted_choices =
@@ -197,7 +341,7 @@ let callback (req: Http_types.request) outchan =
                                id_RE, id]
                               line
                           in
-                          output_string outchan processed_line)
+                          output_string outchan (processed_line ^ "\n"))
                         interactive_user_uri_choice_TPL;
                       raise Chat_unfinished))
 
@@ -225,12 +369,14 @@ let callback (req: Http_types.request) outchan =
                           (fun (id, value) ->
                             (sprintf "alias %s %s"
                               id
-                              (Disambiguate.wrong_xpointer_format_from_wrong_xpointer_format'
+                              (MQueryMisc.wrong_xpointer_format_from_wrong_xpointer_format'
                                 value)))
                           l)) ^
                       "\"")
                     interpretations)
               in
+              Http_daemon.send_basic_headers ~code:200 outchan ;
+              Http_daemon.send_CRLF outchan ;
               iter_file
                 (fun line ->
                   let processed_line =
@@ -239,7 +385,7 @@ let callback (req: Http_types.request) outchan =
                        interpretations_labels_RE, html_interpretations_labels]
                       line
                   in
-                  output_string outchan processed_line)
+                  output_string outchan (processed_line ^ "\n"))
                 interactive_interpretation_choice_TPL;
               raise Chat_unfinished
 
@@ -255,10 +401,10 @@ let callback (req: Http_types.request) outchan =
         in
         (match metasenv' with
         | [] ->
-            let must = MQueryLevels2.get_constraints term' in
-            let must',only = refine_constraints must in
+            let must',only = get_constraints term' req#path in
             let results = MQueryGenerator.searchPattern must' only in 
-            debug_print "FASE 3";
+            Http_daemon.send_basic_headers ~code:200 outchan ;
+            Http_daemon.send_CRLF outchan ;
             iter_file
               (fun line ->
                 let new_aliases =
@@ -268,21 +414,24 @@ let callback (req: Http_types.request) outchan =
                         (List.map
                           (fun name ->
                             sprintf "\"alias %s cic:%s\""
-                              name
+                              (match name with
+                                  CicTextualParser0.Id name -> name
+                                | _ -> assert false (*CSC: completare *))
                               (match f name with
                               | None -> assert false
-                              | Some t ->
-                                  Disambiguate.string_of_cic_textual_parser_uri
-                                    t))
+                              | Some (CicTextualParser0.Uri t) ->
+                                  MQueryMisc.string_of_cic_textual_parser_uri
+                                    t
+                              | _ -> assert false (*CSC: completare *)))
                           domain)
                 in
                 let processed_line =
                   apply_substs
-                    [results_RE, MQueryUtil.text_of_result results "\n";
+                    [results_RE, theory_of_result results ;
                      new_aliases_RE, new_aliases]
                     line
                 in
-                output_string outchan processed_line)
+                output_string outchan (processed_line ^ "\n"))
               final_results_TPL
         | _ -> (* unable to instantiate some implicit variable *)
             Http_daemon.respond
@@ -292,6 +441,8 @@ let callback (req: Http_types.request) outchan =
 
     | invalid_request ->
         Http_daemon.respond_error ~status:(`Client_error `Bad_request) outchan);
+    if req#path <> "/getpage" then
+      MQI.close mqi_options;
     debug_print (sprintf "%s done!" req#path)
   with
   | Chat_unfinished -> prerr_endline "Chat unfinished, Try again!"
@@ -300,15 +451,13 @@ let callback (req: Http_types.request) outchan =
   | exc ->
       Http_daemon.respond
         ~body:(pp_error ("Uncaught exception: " ^ (Printexc.to_string exc)))
-       outchan
+        outchan
 in
 printf "%s started and listening on port %d\n" daemon_name port;
-printf "current directory is %s\n" (Sys.getcwd ());
+printf "Current directory is %s\n" (Sys.getcwd ());
+printf "HTML directory is %s\n" pages_dir;
 flush stdout;
 Unix.putenv "http_proxy" "";
-Mqint.set_database Mqint.postgres_db;
-Mqint.init "host=mowgli.cs.unibo.it dbname=helm_mowgli_new_schema user=helm";
 Http_daemon.start' ~port callback;
-Mqint.close ();
 printf "%s is terminating, bye!\n" daemon_name