Skip to content

H
Heritage_Platform
Commit 36d69f4e authored by Gérard Huet's avatar Gérard Huet
Browse files
Options

Uoh_interface replaced by Scl_parser

parent 67d8d3db
Hide whitespace changes
Inline Side-by-side
Showing with 128 additions and 335 deletions
ML/.depend
View file @ 36d69f4e
......@@ -179,14 +179,14 @@ rank.cmo : ../ZEN/word.cmo web.cmo phases.cmo morphology.cmi \
../ZEN/list2.cmo lexer.cmi constraints.cmi
rank.cmx : ../ZEN/word.cmx web.cmx phases.cmx morphology.cmi \
../ZEN/list2.cmx lexer.cmx constraints.cmx
uoh_interface.cmo : ../ZEN/word.cmo web.cmo SCLpaths.cmo phases.cmo \
morphology.cmi load_morphs.cmo html.cmo dispatcher.cmi
uoh_interface.cmx : ../ZEN/word.cmx web.cmx SCLpaths.cmx phases.cmx \
morphology.cmi load_morphs.cmx html.cmx dispatcher.cmx
reader.cmo : web.cmo uoh_interface.cmo sanskrit.cmi rank.cmo phases.cmo \
scl_parser.cmo : ../ZEN/word.cmo web.cmo SCLpaths.cmo phases.cmo \
morphology.cmi lexer.cmi html.cmo dispatcher.cmi
scl_parser.cmx : ../ZEN/word.cmx web.cmx SCLpaths.cmx phases.cmx \
morphology.cmi lexer.cmx html.cmx dispatcher.cmx
reader.cmo : web.cmo scl_parser.cmo sanskrit.cmi rank.cmo phases.cmo \
paths.cmo html.cmo encode.cmo control.cmo checkpoints.cmo cgi.cmo \
canon.cmo
reader.cmx : web.cmx uoh_interface.cmx sanskrit.cmx rank.cmx phases.cmx \
reader.cmx : web.cmx scl_parser.cmx sanskrit.cmx rank.cmx phases.cmx \
paths.cmx html.cmx encode.cmx control.cmx checkpoints.cmx cgi.cmx \
canon.cmx
parser.cmo : ../ZEN/word.cmo web.cmo uoh_interface.cmo skt_morph.cmi \
......
ML/Makefile
View file @ 36d69f4e
......@@ -5,7 +5,7 @@
# Gérard Huet & Pawan Goyal #
# #
############################################################################
# Makefile of Sanskrit Heritage Software 02-05-2017 Copyright INRIA 2017 #
# Makefile of Sanskrit Heritage Software 02-06-2017 Copyright INRIA 2017 #
############################################################################
# Prerequisites: Ocaml and Camlp4 preprocessor
......@@ -42,9 +42,9 @@ inflected.mli inflected.ml sandhi.ml sandhier.ml pada.ml nouns.mli nouns.ml \
verbs.mli verbs.ml parts.ml conj_infos.mli morpho_string.ml morpho.ml \
declension.ml conjugation.ml indexer.ml indexerd.ml phases.ml lemmatizer.ml \
auto.mli load_transducers.ml dispatcher.mli dispatcher.ml segmenter.ml \
load_morphs.ml lexer.mli lexer.ml rank.ml uoh_interface.ml \
load_morphs.ml lexer.mli lexer.ml rank.ml scl_parser.ml \
reader.ml parser.ml constraints.mli constraints.ml multilingual.ml \
paraphrase.mli paraphrase.ml reader_plugin.ml bank_lexer.ml regression.ml \
paraphrase.mli paraphrase.ml bank_lexer.ml regression.ml \
checkpoints.ml graph_segmenter.ml automaton.ml interface.mli interface.ml \
user_aid.ml reset_caches.ml
......@@ -76,7 +76,8 @@ LINK=ocamlopt -I $(ZEN) -I +camlp4 dynlink.cmxa camlp4lib.cmxa
# standard installation of Sanskrit Heritage platform - assumes ZEN library
engine: test_version cgis reset_caches static_pages regression reader_plugin # parse_apte
engine: test_version cgis reset_caches static_pages regression
# reader_plugin parse_apte
# testing consistency of Heritage_resources and Heritage_platform
test_version: paths.cmx gen.cmx version.cmx control.cmx test_stamp.cmx
......@@ -85,7 +86,7 @@ test_version: paths.cmx gen.cmx version.cmx control.cmx test_stamp.cmx
static_pages: css mk_index_page mk_grammar_page mk_sandhi_page mk_reader_page
all: engine static_pages reset_caches reader_plugin sandhi_test
all: engine static_pages reset_caches sandhi_test
# legacy in need of re-design
regression: rank.cmx regression.cmx
......@@ -184,16 +185,7 @@ inflected.cmx html.cmx SCLpaths.cmx web.cmx naming.cmx morpho_string.cmx morpho.
load_transducers.cmx pada.cmx phases.cmx dispatcher.cmx order.cmx \
chapters.cmx morpho_html.cmx cgi.cmx segmenter.cmx morpho_ext.cmx \
load_morphs.cmx lexer.cmx constraints.cmx rank.cmx bank_lexer.cmx \
uoh_interface.cmx checkpoints.cmx reader.cmx -o reader
reader_plugin: reader_plugin.cmx
$(LINK) unix.cmxa list2.cmx gen.cmx paths.cmx version.cmx date.cmx \
control.cmx word.cmx canon.cmx zen_lexer.cmx phonetics.cmx transduction.cmx \
encode.cmx skt_lexer.cmx padapatha.cmx sanskrit.cmx deco.cmx lexmap.cmx \
inflected.cmx html.cmx SCLpaths.cmx web.cmx naming.cmx morpho_string.cmx morpho.cmx \
load_transducers.cmx pada.cmx phases.cmx dispatcher.cmx order.cmx \
chapters.cmx morpho_html.cmx cgi.cmx segmenter.cmx morpho_ext.cmx \
load_morphs.cmx lexer.cmx constraints.cmx reader_plugin.cmx -o reader_plugin
scl_parser.cmx checkpoints.cmx reader.cmx -o reader
interface: interface.cmx
$(LINK) unix.cmxa list2.cmx gen.cmx paths.cmx version.cmx date.cmx \
......@@ -221,7 +213,7 @@ inflected.cmx html.cmx SCLpaths.cmx web.cmx naming.cmx morpho_string.cmx morpho.
load_transducers.cmx pada.cmx phases.cmx dispatcher.cmx order.cmx \
chapters.cmx morpho_html.cmx bank_lexer.cmx cgi.cmx segmenter.cmx \
morpho_ext.cmx load_morphs.cmx lexer.cmx constraints.cmx checkpoints.cmx \
paraphrase.cmx uoh_interface.cmx parser.cmx -o parser
paraphrase.cmx scl_parser.cmx parser.cmx -o parser
tagger: tagger.cmx
$(LINK) list2.cmx gen.cmx paths.cmx version.cmx date.cmx install.cmx \
......@@ -324,7 +316,7 @@ clean:
rm -f *.cmo *.cmi *.cmx *.ppi *.ppo *.o
rm -f css indexer indexerd sandhier reader parser tagger lemmatizer \
declension conjugation mk_index_page mk_grammar_page mk_reader_page regression \
mk_sandhi_page sandhi_test reader_plugin reset_caches interface user_aid \
mk_sandhi_page sandhi_test reset_caches interface user_aid \
parse_apte tag_apte
# make dico.cmi auto.cmi conj_infos.cmi # needed to get dependencies right
......
ML/install.sv.ml deleted 100644 → 0
View file @ 67d8d3db
(**************************************************************************)
(* *)
(* The Sanskrit Heritage Platform *)
(* *)
(* Gérard Huet *)
(* *)
(* ©2017 Institut National de Recherche en Informatique et en Automatique *)
(**************************************************************************)
(* Module Install reads localisation parameters from paths.ml,
created by "make configure" in main directory, called by configure script.
Describes all installation parameters and resources *)
(*i module Install = struct i*)
(* Heritage used as the generative lexicon *)
value lexicon = "Heritage" (* for lexicon version in Parser.stamp *)
;
(* Configuration of the platform *)
(* truncation is the maximum number of solutions computed by the lexer.
Too small a truncation limit will miss solutions, too large a truncation
limit will provoke un unrecoverable choking server failure. This is relevant
only for the parser (deprecated) mode. The graph interface has no limit. *)
value truncation = 10000
;
(*i end; i*)
ML/interface.ml
View file @ 36d69f4e
......@@ -21,7 +21,7 @@ open Phases; (* [Phases] *)
open Phases; (* [phase is_cache generative] *)
open Dispatcher; (* [transducer_vect phase Dispatch transition trim_tags] *)
open Html;
open Web; (* ps pl abort etc. *)
open Web; (* [ps pl abort reader_cgi] etc. *)
open Cgi;
module Prel = struct (* Interface's lexer prelude *)
......@@ -431,30 +431,24 @@ value check_sentence translit us text_orig checkpoints sentence
; pl (table_begin Spacing20)
; pl tr_begin
; ps (td_wrap (call_undo text checkpoints ^ "Undo"))
; let invoke_web_services n = (* call SCL and SL services *)
if scl_toggle then do
{ if (not iterate.val) && (List.length chunks = 1) then
ps (td_wrap (call_reader text cpts "n" ^ "UoH Nyaya Analysis"))
else ps (td_wrap (call_reader text cpts "o" ^ "UoH Analysis Mode"))
; match corpus with
[ "" -> ()
| id -> invoke_SL sentence cpts id n sent_id link_num
]
}
else () (* these services are not visible unless toggle is set *) in
match count with
[ Num.Int n -> if n=1 (* Unique remaining solution *) then do
{ ps (td_wrap (call_parser text cpts ^ "Unique Solution"))
; invoke_web_services 1
}
else if n < max_count then do
; let call_scl_parser n = (* invocation of scl parser *)
if scl_toggle then
ps (td_wrap (call_reader text cpts "o" ^ "UoH Analysis Mode"))
else () (* [scl_parser] is not visible unless toggle is set *) in
match count with
[ Num.Int n -> if n > max_count then
(* too many solutions would choke the parsers *)
ps (td_wrap ("(" ^ string_of_int n ^ " Solutions)"))
else if n=1 (* Unique remaining solution *) then do
{ ps (td_wrap (call_parser text cpts ^ "Unique Solution"))
; call_scl_parser 1
}
else do
{ ps (td_wrap (call_reader text cpts "p" ^ "Filtered Solutions"))
; let info = string_of_int n ^ if flag then "" else " Partial" in
ps (td_wrap (call_reader text cpts "t" ^ "All " ^ info ^ " Solutions"))
; invoke_web_services n
; call_scl_parser n
}
else (* too many solutions would choke the reader service *)
ps (td_wrap ("(" ^ string_of_int n ^ " Solutions)"))
| _ -> ps (td_wrap "(More than 2^32 Solutions!)")
]
; pl tr_end
......
ML/lexer.ml
View file @ 36d69f4e
......@@ -240,7 +240,7 @@ value print_segment offset (phase,rword,transition) = do
}
;
(* Similarly for [Reader_plugin] mode (without offset) *)
value print_ext_segment counter (phase,rword,_) =
value print_ext_segment counter (phase,rword) =
let print_pada rword =
let word = Morpho_html.visargify rword in
let ic = string_of_int counter in
......@@ -266,7 +266,7 @@ value print_ext_segment counter (phase,rword,_) =
] in
print_ext_tags [] taddhita_phase word sfx_tags
]
; ps "'>"
; ps "'>" (* closes <input *)
; let word = Morpho_html.visargify rword in
ps (Canon.unidevcode word)
; ps td_end
......
ML/lexer.mli
View file @ 36d69f4e
......@@ -52,7 +52,7 @@ module Lexer : functor (* takes its prelude and iterator control as parameters *
value all_checks : ref (list Viccheda.check);
value un_analyzable : Word.word -> (list Disp.segment * Viccheda.resumption);
value set_offset : (int * list Viccheda.check) -> unit;
value print_ext_segment : int -> Disp.segment -> int;
value print_ext_segment : int -> (Phases.phase * Word.word) -> int;
value record_tagging : bool -> bool -> string -> int -> string ->
list (Phases.phase * Word.word * 'a) -> list (int * int) -> unit;
value return_tagging :
......
ML/parser.ml
View file @ 36d69f4e
......@@ -20,7 +20,7 @@ open Html;
open Web; (* ps pl abort truncation etc. [remote_server_host] *)
open Cgi; (* get *)
open Checkpoints;
open Uoh_interface; (* Interface with UoH dependency parser *)
open Scl_parser; (* Interface with UoH dependency parser *)
module Prel = struct (* Parser's lexer prelude *)
......@@ -52,8 +52,6 @@ end (* [Lexer_control] *)
module Lex = Lexer.Lexer Prel Lexer_control
(* [print_proj print_segment_roles print_ext_segment extract_lemma] *)
;
module Ext = UOH Lex
;
value rpc = remote_server_host
and remote = ref False (* local invocation of cgi by default *)
;
......@@ -117,8 +115,9 @@ value analyse query output =
^ ";p=','" ^ string_of_int (find_len top_groups) ^ "')" )
] ^ html_break)
; pl (xml_empty "p")
; if scl_toggle then (* Call SCL parser *)
Ext.print_ext [ (1,List.rev output) ]
; if scl_toggle then (* Call SCL parser *)
let segments = List.map (fun (ph,w,_) -> (ph,w)) output in
Scl_parser.print_scl [ List.rev segments ]
else ()
(*i DEBUG ; Sys.command "ls -l > /tmp/SKT_TEMP/junk" i*)
; List.iter print_bucket top_groups
......
ML/reader.ml
View file @ 36d69f4e
......@@ -31,10 +31,7 @@ open Web; (* ps pl abort etc. [remote_server_host] *)
open Cgi; (* [get decode_url] *)
open Phases; (* [Phases] *)
open Rank; (* [Prel Lex segment_all iterate] *)
open Uoh_interface; (* Interface with UoH dependency parser *)
module Ext = UOH Lex (* [print_ext print_nn] *)
;
(* Reader interface *)
(* Mode parameter of the reader. Controled by service Reader for respectively
tagging, shallow parsing, or dependency analysis with the UoH parser. *)
......@@ -129,19 +126,22 @@ value display limit mode text saved = fun
}
]
}
| Analyse -> match saved with
[ [] -> Ext.print_ext (best_sols:list (int * list Rank.Lex.Disp.segment))
| [ (_,min_buck) :: _ ] ->
let zero_pen = List.append best_sols (List.rev min_buck) in
Ext.print_ext zero_pen
]
| Analyse -> (* [best_sols: list (int * list Rank.Lex.Disp.segment)] *)
let solutions = match saved with
[ [] -> best_sols
| [ (_,min_buck) :: _ ] -> List.append best_sols (List.rev min_buck)
] in
let forget_transitions (phase,word,_) = (phase,word) in
let forget_index (_,segments) = List.map forget_transitions segments in
let segmentations = List.map forget_index best_sols in
Scl_parser.print_scl segmentations
| _ -> ()
]
}
]
;
(* NB This reader is parametrized by an encoding function, that parses the
(* NB This reader is parameterized by an encoding function, that parses the
input as a list of words, according to various transliteration schemes.
However, the use of "decode" below to compute the romanisation and devanagari
renderings does a conversion through VH transliteration which may not be
......
ML/reader_plugin.ml deleted 100755 → 0
View file @ 67d8d3db
(**************************************************************************)
(* *)
(* The Sanskrit Heritage Platform *)
(* *)
(* Gérard Huet & Amba Kulkarni *)
(* *)
(* ©2017 Institut National de Recherche en Informatique et en Automatique *)
(**************************************************************************)
(* This is an adaptation of module Reader from Skt Heritage engine for
external call as a plug-in. It prints on [stdout] an html document giving
segmentation/tagging of its input.
It computes a penalty of the various solutions, and returns all solutions
with minimal penalties (with a further preference for the solutions
having a minimum number of segments), using Constraints for ranking. *)
(*i module Reader_plugin = struct i*)
open Encode;
open Canon;
open Html;
open Web; (* ps, pl, etc. abort truncation *)
open Cgi;
open Morphology;
open Phases;
module Prelude = struct
value prelude () = ()
;
end (* Prelude *)
;
value iterate = ref True (* by default a chunk is a list of words *)
and complete = ref True (* by default we call the fuller segmenter *)
;
module Lexer_control = struct
value star = iterate;
value full = complete;
value out_chan = ref stdout; (* cgi writes on standard output channel *)
end (* [Lexer_control] *)
;
module Lex = Lexer.Lexer Prelude Lexer_control
;
type mode = [ Tag | Parse | Analyse ] (* Segmentation is now obsolete *)
;
(* Builds the penalty stack, grouping together equi-penalty items. *)
(* Beware, [make_groups] reverses the list of tags. *)
value make_groups tagger = comp_rec 1 []
where rec comp_rec seg stack = fun (* going forward in time *)
[ [] -> stack (* result goes backward in time *)
| [ (phase,rword,_) :: rest ] -> (* we ignore euphony transition *)
let word = List.rev rword in
let keep = let tags = tagger phase word in
[ Constraints.roles_of seg word tags :: stack ] in
comp_rec (seg+1) keep rest
]
;
(* Computes minimum penalty in Parse mode *)
value minimum_penalty output =
let tagger = Lex.extract_lemma in
let out = List.rev output in
let groups = make_groups tagger out in
if groups=[] then failwith "Empty penalty stack !" else
let sort_groups = Constraints.sort_flatten groups in
let min_pen = match sort_groups with
[ [] -> failwith "Empty penalty stack"
| [ (pen,_) :: _ ] -> pen
] in
min_pen
;
(* Same as [UoH_interface.print_ext_output] *)
value print_offline_output cho (n,output) =
let ps = output_string cho in
let print_segment = Lex.print_ext_segment in do
{ ps (xml_begin_with_att "solution" [ ("num", string_of_int n) ])
; ps "\n"
; let _ = List.fold_left Lex.print_ext_segment 1 (List.rev output) in ()
; ps (xml_end "solution")
; ps "\n"
}
;
value print_offline_solutions cho =
List.iter (print_offline_output cho)
;
(* Experimental segmentation plug-in of Amba Kulkarni's parser at UoH *)
(* Printing all segmentations on stdout *)
value print_offline = print_offline_solutions stdout
;
(* Compound minimum path penalty with solution length *)
value process_output mode ((_,output) as sol) =
let min = minimum_penalty output in
let m = Constraints.eval_penalty min in
let length_penalty = List.length output in
((m+length_penalty,m),sol)
;
type tagging = (Phases.phase * Word.word * Lex.Disp.transition)
and solution = list tagging
and ranked_solution = (int (* rank *) * solution)
and bucket = (int (* length *) * list ranked_solution)
;
exception No_solution of Word.word
;
exception Solutions of option int and list ranked_solution and list bucket
(* [Solutions None sols saved] returns solutions sols within truncation limit
[Solutions (Some n) sols saved] returns solutions sols within total n
saved is the list of solutions of penalty 0 and worse length penalty *)
;
(* insert builds a triple [(p, sols, saved)] where sols is the list of all pairs
[(m,sol)] such that ranked [sol] has minimal length penalty [p] and
absolute penalty [m] and [saved] is the list of all ranked sols of length
penalty $>$ [p] and absolute penalty 0, arranged in buckets by increasing
length penalty *)
value insert ((pen,min),sol) ((min_pen,sols,saved) as current) =
if sols = [] then (pen,[ (min,sol) ],[])
else if pen>min_pen then if min>0 then current (* sol is thrown away *)
else (min_pen,sols,List2.in_bucket pen sol saved)
else if pen=min_pen then (min_pen,[ (min,sol) :: sols ],saved)
else (pen,[ (min,sol) ],
let rescue = List.fold_right save sols [] in
if rescue=[] then saved else [ (min_pen,rescue) :: saved ])
where save (min,sol) rescued = if min=0 then [ sol :: rescued ]
else rescued
;
(* forget absolute penalties of solutions with minimal length penalty *)
value trim = List.map snd
;
(* does depth-first search in a stack of type [list (output * resumption)] *)
value dove_tail mode init =
dtrec 1 (0,[],[]) init
where rec dtrec n kept stack = (* invariant |stack|=|init| *)
if n > truncation then
let (_,sols,saved) = kept in
raise (Solutions None (trim sols) saved)
else do
{ let total_output = List.fold_right conc stack []
where conc (o,_) oo = o @ oo in
let pen_sol = process_output mode (n,total_output) in
let kept_sols = insert pen_sol kept in
dtrec (n+1) kept_sols (crank [] init stack)
where rec crank acc ini = fun
[ [ (_,c) :: cc ] -> match Lex.Viccheda.continue c with
[ Some next -> List2.unstack acc [ next :: cc ]
| None -> match ini with
[ [ i :: ii ] -> crank [ i :: acc ] ii cc
| _ -> raise (Control.Anomaly "Plugin dove_tail")
(* impossible by invariant *)
]
]
| [] -> let (_,sols,saved) = kept_sols in
raise (Solutions (Some n) (trim sols) saved)
]
}
;
value segment_all mode chunks =
let segs = List.fold_left init [] chunks
where init stack chunk =
let ini_cont = Lex.Viccheda.init_segment chunk in
match Lex.Viccheda.continue ini_cont with
[ Some c -> [ c :: stack ]
| None -> raise (No_solution chunk)
] in
dove_tail mode segs
;
value display limit mode text saved = fun
[ [] -> ()
| best_sols ->
let zero_pen = match saved with
[ [] -> best_sols
| [ (_,min_buck) :: _ ] -> List.append best_sols (List.rev min_buck)
] in
print_offline zero_pen
]
;
value process_sentence text us mode topic (sentence : string) encode =
let chunker = if us then Sanskrit.read_raw_sanskrit
else Sanskrit.read_sanskrit in
let chunks = chunker encode sentence in do
{ let all_chunks = match topic with
[ Some topic -> chunks @ [ code_string topic ]
| None -> chunks
] in
try segment_all mode all_chunks with
[ Solutions limit revsols saved -> let sols = List.rev revsols in
let _ = display limit mode text saved sols in True
| No_solution chunk -> False
]
}
;
value encode = Encode.switch_code "WX" (* encoding in WX as a normalized word *)
;
(* adapt with abort function : string -> string -> unit *)
value abort m1 m2 = raise (Failure (m1 ^ m2))
;
(* input: string is the text to be segmented/parsed *)
(* unsandhied: bool is True is input is unsandhied False if it is sandhied *)
(* topic is (Some "sa.h") (Some "saa") (Some "tat") or None if no topic *)
(* st:bool is True if stemmer for one word, False for tagging sentence *)
(* cp:bool is True if Complete mode, False for Simplified mode *)
value reader_engine input unsandhied topic st cp = do
{ Prelude.prelude ()
; if st then iterate.val:=False else () (* word stemmer *)
; if cp then complete.val:=True else () (* complete reader *)
(* Contextual information from past discourse *)
; try process_sentence "" unsandhied Analyse topic input encode
(* possibly use the returned bool value (success) in your control *)
with [ Stream.Error _ -> abort "Illegal transliteration " input ]
}
;
value safe_engine input unsandhied topic st cp =
try reader_engine input unsandhied topic st cp with
[ Sys_error s -> abort Control.sys_err_mess s (* file pb *)
| Stream.Error s -> abort Control.stream_err_mess s (* file pb *)
| Encode.In_error s -> abort "Wrong input " s
| Exit (* Sanskrit *) -> abort "Wrong character in input - " "use ASCII"
| Invalid_argument s -> abort Control.fatal_err_mess s (* sub *)
| Failure s -> abort Control.fatal_err_mess s (* anomaly *)
| End_of_file -> abort Control.fatal_err_mess "EOF" (* EOF *)
| Not_found (* assoc *) -> abort Control.fatal_err_mess "assoc" (* anomaly *)
| Control.Fatal s -> abort Control.fatal_err_mess s (* anomaly *)
| Control.Anomaly s -> abort Control.fatal_err_mess ("Anomaly: " ^ s)
| _ -> abort Control.fatal_err_mess "Unexpected anomaly"
]
;
(* call [safe_engine input unsandhied topic st cp] with proper parameters *)
let input = input_line stdin in
safe_engine input False None False True
;
(* eg. rAmovanaMgacCawi -> 3 solutions; second is good *)
(*i end; i*)
ML/scl_parser.ml 0 → 100644
View file @ 36d69f4e
(**************************************************************************)
(* *)
(* The Sanskrit Heritage Platform *)
(* *)
(* Gérard Huet & Amba Kulkarni *)
(* *)
(* ©2017 Institut National de Recherche en Informatique et en Automatique *)
(**************************************************************************)
(* Interface with UoH dependency parser *)
open Html;
open Web; (* ps pl etc. *)
open Morphology; (* inflected lemma morphology *)
open Phases; (* Phases *)
open Dispatcher; (* Dispatch *)
open SCLpaths; (* [scl_url scl_cgi] *)
module Prel = struct
value prelude () = Web.reader_prelude Web.reader_title;
end (* Prel *)
;
(* Global parameters of the lexer *)
value iterate = ref True (* by default a chunk is a list of words *)
and complete = ref True (* by default we call the fuller segmenter *)
and output_channel = ref stdout (* by default cgi output on standard output *)
;
module Lexer_control = struct
value star = iterate;
value full = complete;
value out_chan = output_channel
;
end (* [Lexer_control] *)
;
(* Multi-phase lexer *)
module Lex = Lexer.Lexer Prel Lexer_control (* [print_ext] *)
;
value print_scl_segment = Lex.print_ext_segment
;
value print_scl_output output =
List.fold_left print_scl_segment 1 (List.rev output)
;
value print_scl_solutions s =
let _ = print_scl_output s in ()
;
(* Invocation of UoH's CSL parser for dependency graph display *)
value print_scl1 (solutions : list (Phases.phase * Word.word)) =
let svg_interface_url = scl_cgi ^ "SHMT/" in do
{ ps ("<script type=\"text/javascript\" src=\"" ^ scl_url ^ "js_files/dragtable.js\"></script>")
; ps ("<form name=\"word-order\" method=\"POST\" action = \""
^ svg_interface_url ^ "prog/Word_order/call_heritage2anu.cgi\">\n")
; ps ("<table class=\"draggable\">")
; ps tr_begin
; print_scl_solutions solutions
; ps tr_end
; ps table_end
; ps (submit_input "Submit")
}
;
(* We restrict to the first solution - TEMPORARY *)
value print_scl sols = match sols with
[ [] -> failwith "No sol"
| [ s :: _ ] -> print_scl1 s
]
;
(* end; *)
ML/uoh_interface.ml
View file @ 36d69f4e
......@@ -54,22 +54,21 @@ module UOH
(* Interface with Amba Kulkarni's parser at UoH - Analysis mode *)
(****************************************************************)
value print_ext_output (_,output) =
List.fold_left Lex.print_ext_segment 1 (List.rev output)
;
(* Delimitor for offline printing and piping into UoH's parser *)
(* UNUSED Delimitor for offline printing and piping into UoH's parser
value delimitor = fun
[ Iic | Iic2 | A | An | Iicv | Iicc | Iik | Iikv | Iikc | Iiif | Iiy -> "-"
| Iiv | Iivv | Iivc -> "++"
| Pv | Pvk | Pvkc | Pvkv -> failwith "No more Pv segments"
| _ -> " "
]
; *)
value print_ext_output (_,output) =
List.fold_left Lex.print_ext_segment 1 (List.rev output)
;
value print_ext_solutions s =
let _ = print_ext_output s in ()
;
(* Invocation of UoH's CSL parser for dependency graph display *)
value print_ext1 (solutions : (int * list Lex.Disp.segment)) = do
{ ps ("<script type=\"text/javascript\" src=\"" ^ scl_url ^ "js_files/dragtable.js\"></script>")
......@@ -83,7 +82,8 @@ value print_ext1 (solutions : (int * list Lex.Disp.segment)) = do
; ps (submit_input "Submit")
}
;
value print_ext sols = match sols with
(* We restrict to the first solution - TEMPORARY *)
value print_ext sols = match sols with
[ [] -> failwith "No sol"
| [ s :: _ ] -> print_ext1 s
]
......
ML/version.ml
View file @ 36d69f4e
......@@ -8,4 +8,4 @@
(**************************************************************************)
(* Generated by make version - see main Makefile *)
value version="2.99" and version_date="2017-06-01";
value version="2.99" and version_date="2017-06-02";
SETUP/MMakefile
View file @ 36d69f4e
......@@ -46,6 +46,7 @@ SHELL=/bin/sh
# LGPL distribution Kerala workshop 25-12-2011 Happy Christmas !
# LGPL distribution WSC Conference Bangkok 25-06-2015
# GitLab archive 15-04-2017
# SCL parser interface 02-06-2017
SOFT_LICENSE=LEGAL/LGPL/LICENSE # Free software license LGPL
#SOFT_LICENSE=LEGAL/UoH/LICENSE # University of Hyderabad distribution
......@@ -226,9 +227,9 @@ ML/verbs.mli ML/verbs.ml ML/parts.ml ML/morpho.ml ML/declension.ml \
ML/conjugation.ml ML/indexer.ml ML/indexerd.ml ML/phases.ml ML/lemmatizer.ml \
ML/auto.mli ML/load_transducers.ml ML/dispatcher.mli ML/dispatcher.ml \
ML/morpho_string.ml ML/segmenter.ml ML/load_morphs.ml ML/lexer.mli ML/lexer.ml \
ML/rank.ml ML/uoh_interface.ml ML/reader.ml ML/parser.ml ML/constraints.mli \
ML/rank.ml ML/scl_parser.ml ML/reader.ml ML/parser.ml ML/constraints.mli \
ML/constraints.ml ML/multilingual.ml ML/paraphrase.mli ML/paraphrase.ml \
ML/reader_plugin.ml ML/bank_lexer.ml ML/regression.ml ML/checkpoints.ml \
ML/bank_lexer.ml ML/regression.ml ML/checkpoints.ml \
ML/graph_segmenter.ml ML/interface.mli ML/interface.ml ML/user_aid.ml \
ML/reset_caches.ml
......
SETUP/version.txt
View file @ 36d69f4e
VERSION='2.99'
DATE='2017-06-01'
DATE='2017-06-02'
SITE/manual.html
View file @ 36d69f4e
......@@ -833,7 +833,7 @@ is available as a text document README in the distribution directory.
The complete Ocaml source of all modules of the Heritage Engine is available
in literate programming style as a pdf document
<a href="DOC/Heritage_platform.pdf">
<strong>Heritage_engine_documentation</strong></a>.
<strong>Heritage_platform_documentation</strong></a>.
It may be considered as our vyākaraṇasūtrasaṃgraha.
<h2 class="b2" id="installation">How to install the Heritage Engine on your own server</h2>
......
SITE/xml.html
View file @ 36d69f4e
......@@ -84,7 +84,7 @@ preverb sequences, given with their sandhi analysis.
All these linguistic data banks are Copyrighted Gérard Huet 1994-2017.
They are derived from the Sanskrit Heritage Dictionary
version 2.99 dated 2017-06-01.
version 2.99 dated 2017-06-02.
<p>
Use of these linguistic resources is granted according to the
Lesser General Public Licence for Linguistic Resources.
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment