let hash x = 131*x/(x-1+1)
let test_loop tree tag =
- let t' = Tree.tagged_desc tree tag Tree.root in
+ let t' = Tree.tagged_descendant tree tag Tree.root in
let f = Hashtbl.create 4096
in
- let jump t _ = Tree.tagged_foll_ctx tree tag t Tree.root in
+ let jump t _ = Tree.tagged_following_below tree tag t Tree.root in
let g t ctx =
if t == Tree.nil then 0
else 1+ ((Hashtbl.find f (hash 101)) (jump t ctx) ctx)
let test_loop2 tree tag =
- let t' = Tree.tagged_desc tree tag Tree.root in
+ let t' = Tree.tagged_descendant tree tag Tree.root in
let f = Hashtbl.create 4096
in
- let jump t _ = Tree.tagged_foll_ctx tree tag t Tree.root in
+ let jump t _ = Tree.tagged_following_below tree tag t Tree.root in
let rec g t ctx =
if t == Tree.nil then 0
else 1+ (match (Hashtbl.find f (hash 101)) with
Hashtbl.add f (hash 101) `Foo;
g t' Tree.root
+let test_text doc =
+ let _ = Printf.eprintf "Contains(bree)" in
+ let _ = time (Tree.test_contains doc) "bree" in
+ let _ = Printf.eprintf "Contains(brain)" in
+ let _ = time (Tree.test_contains doc) "brain" in
+ let _ = Printf.eprintf "Contains(brain)" in
+ let i = time (Tree.test_contains doc) "brain" in
+ let _ = Printf.eprintf "%i\nContains(Australia)" i in
+ let i = time (Tree.test_contains doc) "AUSTRALIA" in
+ let _ = Printf.eprintf "%i\n Contains(1930)" i in
+ let i = time (Tree.test_contains doc) "1930" in
+ let _ = Printf.eprintf "%i\n startswith(bar)" i in
+ let i = time (Tree.test_prefix doc) "bar" in
+ let _ = Printf.eprintf "%i\n endswith(LAND)" i in
+ let i = time (Tree.test_suffix doc) "LAND" in
+ let _ = Printf.eprintf "%i\n =(2001)" i in
+ let i = time (Tree.test_equals doc) "2001" in
+ let _ = Printf.eprintf "%i\n =(Nguyen)" i in
+ let i = time (Tree.test_equals doc) "Nguyen" in
+ Printf.eprintf "%i\n" i ;
+ ()
+
+type pointers
+external build_pointers : Tree.t -> pointers = "caml_build_pointers"
+external iter_pointers : pointers -> int = "caml_iter_pointers"
+external free_pointers : pointers -> unit = "caml_free_pointers"
+
+
let main v query_string output =
let _ = Tag.init (Tree.tag_pool v) in
with
Ulexer.Loc.Exc_located ((x,y),e) -> Printf.eprintf "character %i-%i %s\n" x y (Printexc.to_string e);exit 1
in
+ let _ = Printf.eprintf "Number of nodes %i\n%!" (Tree.size v) in
+ let _ = test_text v in
+(* let _ = Tree.stats v in
+ let _ = Printf.eprintf "Timing first_child/next_sibling %!" in
+ let _ = time (Tree.benchmark_fcns) v in
+ let _ = Printf.eprintf "Timing last_child/prev_sibling %!" in
+ let _ = time (Tree.benchmark_lcps) v in
+ let _ = Printf.eprintf "Timing jump to a %!" in
+ let _ = time (Tree.benchmark_jump v) (Tag.tag "a") in
+ let _ = Printf.eprintf "Timing pointer allocation %!" in
+ let pointers = time (build_pointers) v in
+ let _ = Printf.eprintf "Timing pointer iteration %!" in
+ let i = time (iter_pointers) pointers in
+ let _ = Printf.eprintf "Traversed %i pointers\n\nTiming pointer deallocation %!" i in
+ let _ = time (free_pointers) pointers in *)
(* let _ = Printf.eprintf "Timing //keyword :" in
let r = time (test_loop v) (Tag.tag "keyword") in
let _ = Printf.eprintf "Count is %i\n%!" r in
let _ = Printf.eprintf "Timing //keyword 2:" in
let r = time (test_loop2 v) (Tag.tag "keyword") in
- let _ = Printf.eprintf "Count is %i\n%!" r in *)
+ let _ = Printf.eprintf "Count is %i\n%!" r in
let _ = Printf.eprintf "Timing //node() :" in
- let _ = time (test_full) v in
+ let _ = time (test_full) v in *)
XPath.Ast.print Format.err_formatter query;
Format.fprintf Format.err_formatter "\n%!";
Printf.eprintf "Compiling query : ";
let jump_to =
match contains with
None -> (max_int,`NOTHING)
- | Some s ->
+ | Some ((op,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;
+ time (Tree.init_textfun op v) s;
end
else begin
Printf.eprintf "Naive contains\nCalling global contains : ";
let _ = Printf.eprintf "%!" 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
+ if !Options.count_only then
+ let r = time_mem (bottom_up_count auto v )(snd test_list) in
let _ = Printf.eprintf "Number of nodes in the result set : %i\n%!" r
in ()
+ else begin
+ let r = time_mem (bottom_up auto v )(snd test_list) in
+ let _ = Printf.eprintf "Number of nodes in the result set : %i\n%!" (IdSet.length r)
+ in
+ match output with
+
+ | None -> ()
+ | Some f ->
+ Printf.eprintf "Serializing results : ";
+ time( fun () ->
+ (*let oc = open_out f in *)
+ let oc = Unix.openfile f [ Unix.O_WRONLY;Unix.O_TRUNC;Unix.O_CREAT] 0o644 in
+ (*output_string oc "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n";*)
+ IdSet.iter (fun t ->
+ Tree.print_xml_fast3 v t oc;
+ (*output_char oc '\n'; *)
+ ) r) ();
+ end
+
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
+ let r = time_mem ( top_down_count auto ) v in
let _ = Printf.eprintf "Number of nodes in the result set : %i\n%!" r
in ()
else
let module GR = Ata.Test(struct let doc = v end) in
- let result = time (GR.top_down auto) v in
+ let result = time_mem (GR.top_down auto) v in
let _ = Printf.eprintf "Counting results " in
let rcount = time (GR.Results.length) result in
Printf.eprintf "Number of nodes in the result set : %i\n" rcount;
(*let oc = open_out f in *)
let oc = Unix.openfile f [ Unix.O_WRONLY;Unix.O_TRUNC;Unix.O_CREAT] 0o644 in
(*output_string oc "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n";*)
+ let t1 = ref (Unix.gettimeofday()) in
+ let count = ref 1 in
+ let old_count = ref 1 in
GR.Results.iter (fun t ->
- Tree.print_xml_fast3 v t oc;
+ incr count;
+ begin
+ if (!count mod 15) == 0
+ then
+ let t2 = Unix.gettimeofday() in
+ let _ = Printf.eprintf "Printing %i elements in %f ms\n"
+ (!count - !old_count) (1000. *.(t2 -. !t1))
+ in
+ ( old_count := !count; t1 := Unix.gettimeofday())
+ end;
+ Tree.print_xml_fast3 v t oc;
(*output_char oc '\n'; *)
- ) result) ();
+ ) result) ();
end;
end;
let _ = Gc.set enabled_gc in
then
begin
Printf.eprintf "Loading from file : ";
- time (Tree.load ~sample:!Options.sample_factor )
+ time (Tree.load ~sample:!Options.sample_factor ~load_text:(not !Options.count_only))
!Options.input_file;
end
else