X-Git-Url: http://git.nguyen.vg/gitweb/?a=blobdiff_plain;f=main.ml;h=4fcc0af9e2ed5631c5b2886e691ffec8ab270b94;hb=70ff0bfc463882ecf233f1b1a7ac4a8007fa4cc2;hp=c8ed40e1cfed841f0c95f99ba0da467a8ab10a13;hpb=f98a8d98d86941a885f492d5cc134e34989c198a;p=SXSI%2Fxpathcomp.git diff --git a/main.ml b/main.ml index c8ed40e..4fcc0af 100644 --- a/main.ml +++ b/main.ml @@ -4,7 +4,6 @@ (* Copyright NICTA 2008 *) (* Distributed under the terms of the LGPL (see LICENCE) *) (******************************************************************************) -INCLUDE "debug.ml" open Ata @@ -24,88 +23,100 @@ let disabled_gc = { Gc.get() with Gc.max_overhead = 1000000; Gc.space_overhead = 100 } -let main v query output = +let main v query_string output = let _ = Tag.init (Tree.tag_pool v) in Printf.eprintf "Parsing query : "; let query = try time - XPath.Parser.parse_string query + XPath.Parser.parse_string query_string with Ulexer.Loc.Exc_located ((x,y),e) -> Printf.eprintf "character %i-%i %s\n" x y (Printexc.to_string e);exit 1 in XPath.Ast.print Format.err_formatter query; Format.fprintf Format.err_formatter "\n%!"; Printf.eprintf "Compiling query : "; - let auto,ltags,contains = time XPath.Compile.compile query in + let auto,ltags,contains = time (XPath.Compile.compile ~querystring:query_string) query in let _ = Ata.dump Format.err_formatter auto in let _ = Printf.eprintf "%!" in - - let do_contains = match contains with - None -> false - | Some s -> - let r = Tree.count v s - in - Printf.eprintf "%i documents in the TextCollection\n" (Tree.text_size v); - Printf.eprintf "Global count is %i, using " r; - if r < !Options.tc_threshold then begin - Printf.eprintf "TextCollection contains\nTiming call to raw global contains (1st time): "; - time (Tree.unsorted_contains v) s; - Printf.eprintf "Calling global contains : "; - time (Tree.init_contains v) s; - Printf.eprintf "Timing call to global count contains : "; - let r = time (Tree.count_contains v) s - in - Printf.eprintf " number of matching nodes %i \n%!" r; - Printf.eprintf "Timing call to raw global contains (2nd time): "; - time (Tree.unsorted_contains v) s; - end - else begin - Printf.eprintf "Naive contains\nCalling global contains : "; - time (Tree.init_naive_contains v) s - end;true + let jump_to = + match contains with + None -> (max_int,`NOTHING) + | Some s -> + let r = Tree.count v s + in + Printf.eprintf "%i documents in the TextCollection\n" (Tree.text_size v); + Printf.eprintf "Global count is %i, using " r; + if r < !Options.tc_threshold then begin + Printf.eprintf "TextCollection contains\nCalling global contains : "; + time (Tree.init_contains v) s; + end + else begin + Printf.eprintf "Naive contains\nCalling global contains : "; + time (Tree.init_naive_contains v) s + end;(r,`CONTAINS(s)) + in + let test_list = + if (!Options.backward) then begin + Printf.eprintf "Finding min occurences : "; + time + ( List.fold_left (fun ((min_occ,kind)as acc) (tag,_) -> + let numtags = Tree.subtree_tags v tag Tree.root in + if ((numtags < min_occ) && numtags >= 2) + then (numtags,`TAG(tag)) + else acc) jump_to) ltags + end + else (max_int,`NOTHING) + in + let _ = if (snd test_list) != `NOTHING then + let occ,s1,s2 = match test_list with + | (x,`TAG (tag)) -> (x, "tag", (Tag.to_string tag)) + | (x,`CONTAINS(s)) -> (x, "contains()", s) + | _ -> assert false + in + Printf.eprintf "Will jump to %s %s occuring %i time\n%!" s1 s2 occ in Printf.eprintf "Execution time %s : " (if !Options.count_only then "(counting only)" else if !Options.backward then "(bottomup)" else ""); begin let _ = Gc.full_major();Gc.compact() in - let _ = Gc.set (disabled_gc) in - if !Options.backward then - let tag,set = List.hd ltags in - let r = if do_contains - then time (bottom_up_count_contains auto) v - else time (bottom_up_count auto v) tag in + (* let _ = Gc.set (disabled_gc) in *) + if !Options.backward && ((snd test_list) != `NOTHING )then + + let r = time (bottom_up_count auto v )(snd test_list) in let _ = Printf.eprintf "Number of nodes in the result set : %i\n%!" r in () else + let _ = + if !Options.backward then Printf.eprintf "WARNING: couldn't find a jumping point, running top-down\n" + in if !Options.count_only then - let r = time ( top_down_count auto ) v in (* not clean *) - let _ = Printf.eprintf "Number of nodes in the result set : %i\n%!" r - in () - else - let result = time (top_down auto) v in - let rcount = IdSet.length result in - Printf.eprintf "Number of nodes in the result set : %i\n" rcount; - Printf.eprintf "\n%!"; - begin - match output with - | None -> () - | Some f -> - Printf.eprintf "Serializing results : "; - time( fun () -> - let oc = open_out f in - output_string oc "\n"; - IdSet.iter (fun t -> - Tree.print_xml_fast oc t; - output_char oc '\n'; - output_string oc "----------\n"; - ) result) (); - end; + let r = time ( top_down_count auto ) v in + let _ = Printf.eprintf "Number of nodes in the result set : %i\n%!" r + in () + else + let result = time (top_down auto) v in + let rcount = IdSet.length result in + Printf.eprintf "Number of nodes in the result set : %i\n" rcount; + Printf.eprintf "\n%!"; + begin + match output with + | None -> () + | Some f -> + Printf.eprintf "Serializing results : "; + time( fun () -> + let oc = open_out f in + output_string oc "\n"; + IdSet.iter (fun t -> + Tree.print_xml_fast oc v t; + output_char oc '\n'; + output_string oc "----------\n"; + ) result) (); + end; end; let _ = Gc.set enabled_gc in -(* let _ = Ata.dump Format.err_formatter auto in *) - Printf.eprintf "Total running time : %fms\n%!" (total_time()) + Printf.eprintf "Total running time : %fms\n%!" (total_time()) ;; - + Options.parse_cmdline();; let v = @@ -131,10 +142,3 @@ let v = in main v !Options.query !Options.output_file;; -IFDEF DEBUG -THEN -Printf.eprintf "\n=================================================\nDEBUGGING\n%!"; - -Tree.DEBUGTREE.print_stats Format.err_formatter;; -Gc.full_major() -ENDIF