(******************************************************************************)
INCLUDE "debug.ml"
-open Automaton
+open Ata
let l = ref [] ;;
let total_time () = List.fold_left (+.) 0. !l;;
-let main filename query output =
- (* Just a trick to allow the C++ code to print debugging stuff first *)
- let v = time (fun () -> let v = Tree.Binary.parse_xml_uri filename;
- in Printf.eprintf "Parsing document : %!";v
- ) ()
- in
+let test_slashslash tree k =
+ let test =
+ match k with "*" -> TagSet.remove (Tag.tag "") TagSet.star
+ | s -> TagSet.singleton (Tag.tag k)
+ in
+ let attorstring = TagSet.cup TagSet.pcdata TagSet.attribute in
+ let rec aux t acc =
+ if Tree.Binary.is_node t
+ then
+ let tag = Tree.Binary.tag t in
+ let l = Tree.Binary.first_child t
+ and r = Tree.Binary.next_sibling t
+ in
+ let acc =
+ if TagSet.mem tag test
+ then
+ TS.append t acc
+ else
+ acc
+ in
+ let rl = if TagSet.mem tag attorstring then acc else aux l acc
+ in aux r rl
+ else
+ acc
+ in
+ let _ = Printf.eprintf "Testing optimal //%s ... " k in
+ let r = time (aux tree ) TS.empty in
+ Printf.eprintf "Result set is %i nodes\n%!" (TS.length r)
+
+
+let test_jump tree k =
+ let ttag = Tag.tag k in
+
+ let rec loop acc tree =
+ if Tree.Binary.is_node tree
+ then
+ let acc = TS.cons tree acc in
+ loop acc (Tree.Binary.tagged_next tree ttag)
+ else
+ acc
+
+ in
+ let _ = Printf.eprintf "Testing jumping for tag %s ... " k in
+ let r = time (loop TS.empty ) (Tree.Binary.tagged_next tree ttag) in
+ Printf.eprintf "Result set is %i nodes\n%!" (TS.length r)
+
+
+
+let test_traversal tree k =
+ let ttag = Tag.tag k in
+ let iid t = if Tree.Binary.is_node t then Tree.Binary.id t else -1 in
+ let rec aux t =
+ if Tree.Binary.is_node t
+ then
+ let tag = Tree.Binary.tag t in
+ let l = Tree.Binary.first_child t
+ and r = Tree.Binary.next_sibling t
+ in
+ let _ = Printf.eprintf "Tree with id %i and tag=%s, tagged_desc %s is %i tagged_foll is %i, tagged_next is %i\n%!"
+ (Tree.Binary.id t) (Tag.to_string tag) (k)
+ (iid (Tree.Binary.tagged_desc t ttag))
+ (iid (Tree.Binary.tagged_foll t ttag))
+ (iid (Tree.Binary.tagged_next t ttag))
+ in
+ aux l;
+ aux r;
+
+ else
+ ()
+ in
+ aux tree
+
+let test_count_subtree tree k =
+ let ttag = Tag.tag k in
+ let _ = Printf.eprintf "Counting subtrees with tag %s ... %!" k in
+ let r = time(Tree.Binary.subtree_tags tree) ttag in
+ Printf.eprintf "%i nodes \n%!" r
+
+let main v query output =
let _ = Tag.init (Tree.Binary.tag_pool v) in
- MM(v,__LOCATION__);
Printf.eprintf "Parsing query : ";
let query = try
time
XPath.Parser.parse_string query
with
Ulexer.Loc.Exc_located ((x,y),e) -> Printf.eprintf "character %i-%i %s\n" x y (Printexc.to_string e);exit 1
- in
- Printf.eprintf "Compiling query : ";
+ in
+ XPath.Ast.print Format.err_formatter query;
+ Format.fprintf Format.err_formatter "\n%!";
+ Printf.eprintf "Compiling query : ";
let auto = time XPath.Compile.compile query in
- XPath.Ast.print Format.err_formatter query;
- Format.eprintf "\n%!";
- (* Format.eprintf "Internal rep of the tree is :\n%!";
- Tree.Binary.dump v; *)
+
+ let _ = Ata.dump Format.err_formatter auto ;
+ Format.fprintf Format.err_formatter "\n%!"
+ in
+ let _ = test_count_subtree v "keyword" in
+ let _ = test_jump v "keyword" in
+
Printf.eprintf "Execution time : ";
- time (fun v -> ignore (TopDown.accept auto v)) v;
- Printf.eprintf "Number of nodes in the result set : %i\n" (BST.cardinal auto.result);
- begin
- match output with
- | None -> ()
- | Some f ->
-
- Printf.eprintf "Serializing results : ";
- time( fun () ->
- let oc = open_out f in
- output_string oc "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n";
- BST.iter (fun t -> Tree.Binary.print_xml_fast oc t;
- output_char oc '\n') auto.result) ();
- end;
- Printf.eprintf "Total time : %fms\n Coherence : %i\n%!" (total_time())
+ let result = time (BottomUpNew.run auto) v in
+ Printf.eprintf "Number of nodes in the result set : %i\n" (TS.length result);
+ begin
+ match output with
+ | None -> ()
+ | Some f ->
+
+ Printf.eprintf "Serializing results : ";
+ time( fun () ->
+ let oc = open_out f in
+ output_string oc "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n";
+ TS.iter (fun t -> Tree.Binary.print_xml_fast oc t;
+ output_char oc '\n') result) ();
+ end;
+ (* let _ = Ata.dump Format.err_formatter auto in
+ Format.fprintf Format.err_formatter "\n%!"; *)
+ Printf.eprintf "Total time : %fms\n%!" (total_time())
;;
-let argc = Array.length Sys.argv;;
-if (argc < 3 || argc >4)
-then
- (prerr_endline ("usage : " ^ Sys.argv.(0) ^ " <document> \'query\'[ <output> ]");
- exit 1)
-;;
+Options.parse_cmdline();;
-main Sys.argv.(1) Sys.argv.(2) (if argc == 4 then Some Sys.argv.(3) else None) ;;
+let v =
+ if (Filename.check_suffix !Options.input_file ".srx")
+ then
+ begin
+ Printf.eprintf "Loading from file : ";
+ time (Tree.Binary.load ~sample:!Options.sample_factor )
+ (Filename.chop_suffix !Options.input_file ".srx");
+ end
+ else
+ let v =
+ time (fun () -> let v = Tree.Binary.parse_xml_uri !Options.input_file;
+ in Printf.eprintf "Parsing document : %!";v
+ ) ()
+ in
+ if !Options.save_file <> ""
+ then begin
+ Printf.eprintf "Writing file to disk : ";
+ time (Tree.Binary.save v) !Options.save_file;
+ end;
+ v
+in
+ main v !Options.query !Options.output_file;;
+IFDEF DEBUG
+THEN
Printf.eprintf "\n=================================================\nDEBUGGING\n%!";
-Tree.DEBUGTREE.print_stats Format.err_formatter;;
-
+Tree.DEBUGTREE.print_stats Format.err_formatter;;
+Gc.full_major()
+ENDIF