Make tables module-aware; prepare for adding a database

Author: Jan Wielemaker

README.TXT

@@ -45,12 +45,13 @@ unoconv -f ods *.xlsx
 
 ---+ Approach
 
----++ Dependency discovery
+  1. Find clear tables: a square of equally typed objects with
+     aligned strings rows and columns.
 
-Dependency of a cell is a cell or datasource.
+  2. Delete tables that are completely inside other tables. These are
+     typically the header of another table.
 
----++ Identify Datasources
+  3. Find conflicts.  These are tables that intersect.
 
-A datasource is a block of cells (cell_range(Sheet,SX,SY,EX,EY)).
-
-  - Identify as dependency from formulas
+     - Assume this is the result of header rows/columns that are shared.
+     -

datasource.pl

@@ -1,6 +1,7 @@
 :- module(datasource,
 	  [ ds_sheet/2,			% +DS, -Sheet
 	    ds_size/3,			% +DS, -Columns, -Rows
+	    ds_id/2,			% ?DS, ?Id
 
 	    ds_inside/3,		% +DS, ?X, ?Y
 
@@ -14,6 +15,8 @@
 	    ds_column_slice/3,		% +DS1, ?Offset, ?Slice
 	    ds_unbounded_column_slice/3	% +DS1, +Offset, ?Slice
 	  ]).
+:- use_module(ods_table).
+
 
 		 /*******************************
 		 *	 SIMPLE PROPERTIES	*
@@ -33,6 +36,23 @@
 	Columns is EX-SX+1,
 	Rows is EY-SY+1.
 
+%%	ds_id(+DS, -ID) is det.
+%
+%	True when ID is an identifier for DS
+
+ds_id(DS, Id) :-
+	ground(DS), !,
+	DS = cell_range(Sheet, SX,SY, EX,EY),
+	column_name(SX, SC),
+	column_name(EX, EC),
+	(   sheet_name_need_quotes(Sheet)
+	->  format(atom(Id), '[\'~w\'.~w~w:~w~w]', [Sheet,SC,SY,EC,EY])
+	;   format(atom(Id), '[~w.~w~w:~w~w]', [Sheet,SC,SY,EC,EY])
+	).
+ds_id(DS, Id) :-
+	atom_codes(Id, Codes),
+	phrase(ods_reference(DS, ''), Codes).
+
 
 		 /*******************************
 		 *	    COORDINATES		*

ods_table.pl

@@ -15,7 +15,10 @@
 	    cell/8,			% :Sheet, ?X, ?Y, ?V, ?T, ?F, ?S, ?A
 
 	    column_name/2,		% ?Index, ?Name
-	    ods_DOM/3			% :Source, -DOM, +Options
+	    ods_DOM/3,			% :Source, -DOM, +Options
+
+	    sheet_name_need_quotes/1,	% +SheetName
+	    ods_reference//2		% -Expr, +Table
 	  ]).
 :- use_module(library(xpath)).
 :- use_module(library(sgml)).
@@ -831,6 +834,15 @@
 not_in_sheet_name(0'#).
 not_in_sheet_name(0'$).
 
+%%	sheet_name_need_quotes(+Name) is semidet.
+%
+%	True when Name is a sheet name that needs (single) quotes.
+
+sheet_name_need_quotes(Name) :-
+	atom_codes(Name, Codes),
+	member(Code, Codes),
+	not_in_sheet_name(Code), !.
+
 
 		 /*******************************
 		 *	 CELL PROPERTIES	*

recognise.pl

@@ -66,7 +66,7 @@
 	    cell_class(M:Sheet, SX,AY, TAbove), TAbove \== Type
 	).
 
-%%	block(?DataSource, ?Type) is nondet.
+%%	block(:DataSource, ?Type) is nondet.
 
 block(M:cell_range(Sheet, SX,SY, EX,EY), Type) :-
 	block(M:Sheet, SX,SY, EX,EY, Type).

table.pl

@@ -7,6 +7,9 @@
 :- use_module(datasource).
 :- use_module(library(lists)).
 
+:- meta_predicate
+	tables(:, ?, -).
+
 %%	tables(?Sheet, +Type, -Tables) is det.
 %
 %	Make an initial guess  at  all  tables.   Table  is  a  list  of
@@ -22,13 +25,14 @@
 		NestedTables),
 	append(NestedTables, Tables).
 
-table_in_sheet(Sheet, Type, table(DS,Headers,Union)) :-
+table_in_sheet(M:Sheet, Type, table(Id,Type,DS,Headers,Union)) :-
 	ds_sheet(DS, Sheet),
 	cell_class(Type),
 	anchor(DS, Type),
-	once((block(DS, Type),
-	      table(DS, Headers))),
-	ds_union([DS|Headers], Union).
+	once((block(M:DS, Type),
+	      table(M:DS, Headers))),
+	ds_union([DS|Headers], Union),
+	ds_id(DS, Id).
 
 %%	remove_inside(+Tables0, -Tables) is det.
 %
@@ -67,50 +71,55 @@
 		Cells).
 
 
-%%	table(+DataDS, ?SupportDS) is nondet.
+%%	table(:DataDS, ?SupportDS) is nondet.
 %
 %	True when there is a table  with   DataDS  and a list of support
 %	datasources.
 
-table(DataDS, TitleDS) :-
+table(QDataDS, TitleDS) :-
+	QDataDS = _:DataDS,
 	ds_size(DataDS, Cols, Rows),
-	top_rows(DataDS, -1, TitleDS, Left),
-	left_columns(DataDS, -1, Left, Right),
-	right_columns(DataDS, Cols, Right, Bottom),
-	bottom_rows(DataDS, Rows, Bottom, []).
-
-%%	top_rows(+DS, +StartIndex, -Rows, ?Tail) is nondet.
-%%	bottom_rows(+DS, +StartIndex, -Rows, ?Tail) is nondet.
-%%	left_columns(+DS, +StartIndex, -Rows, ?Tail) is nondet.
-%%	right_columns(+DS, +StartIndex, -Rows, ?Tail) is nondet.
-
-top_rows(DataDS, Index, [Row|Rows], Tail) :-
+	top_rows(QDataDS, -1, TitleDS, Left),
+	left_columns(QDataDS, -1, Left, Right),
+	right_columns(QDataDS, Cols, Right, Bottom),
+	bottom_rows(QDataDS, Rows, Bottom, []).
+
+%%	top_rows(:DS, +StartIndex, -Rows, ?Tail) is nondet.
+%%	bottom_rows(:DS, +StartIndex, -Rows, ?Tail) is nondet.
+%%	left_columns(:DS, +StartIndex, -Rows, ?Tail) is nondet.
+%%	right_columns(:DS, +StartIndex, -Rows, ?Tail) is nondet.
+
+top_rows(QDataDS, Index, [Row|Rows], Tail) :-
+	QDataDS = M:DataDS,
 	ds_unbounded_row_slice(DataDS, Index, Row),
-	row(Row, string),
+	row(M:Row, string),
 	Up is Index - 1,
-	top_rows(DataDS, Up, Rows, Tail).
+	top_rows(QDataDS, Up, Rows, Tail).
 top_rows(_, _, Tail, Tail).
 
 
-bottom_rows(DataDS, Index, [Row|Rows], Tail) :-
+bottom_rows(QDataDS, Index, [Row|Rows], Tail) :-
+	QDataDS = M:DataDS,
 	ds_unbounded_row_slice(DataDS, Index, Row),
-	row(Row, string),
+	row(M:Row, string),
 	Down is Index + 1,
-	bottom_rows(DataDS, Down, Rows, Tail).
+	bottom_rows(QDataDS, Down, Rows, Tail).
 bottom_rows(_, _, Tail, Tail).
 
 
-left_columns(DataDS, Index, [Col|Cols], Tail) :-
+left_columns(QDataDS, Index, [Col|Cols], Tail) :-
+	QDataDS = M:DataDS,
 	ds_unbounded_column_slice(DataDS, Index, Col),
-	col(Col, string),
+	col(M:Col, string),
 	Up is Index - 1,
-	left_columns(DataDS, Up, Cols, Tail).
+	left_columns(QDataDS, Up, Cols, Tail).
 left_columns(_, _, Tail, Tail).
 
-right_columns(DataDS, Index, [Col|Cols], Tail) :-
+right_columns(QDataDS, Index, [Col|Cols], Tail) :-
+	QDataDS = M:DataDS,
 	ds_unbounded_column_slice(DataDS, Index, Col),
-	col(Col, string),
+	col(M:Col, string),
 	Right is Index + 1,
-	right_columns(DataDS, Right, Cols, Tail).
+	right_columns(QDataDS, Right, Cols, Tail).
 right_columns(_, _, Tail, Tail).
 

test.pl

@@ -144,7 +144,11 @@
 	integer(SX), integer(EX),
 	column_name(SX, CS),
 	column_name(EX, CE),
-	format('[~q.~w~w:~w~w]', [Sheet, CS,SY,CE,EY]).
+	(   atom(Sheet),
+	    \+ sheet_name_need_quotes(Sheet)
+	->  format('[~w.~w~w:~w~w]', [Sheet, CS,SY,CE,EY])
+	;   format('[\'~w\'.~w~w:~w~w]', [Sheet, CS,SY,CE,EY])
+	).
 
 user:message_property(debug(ods(test(ok))), color(fg(green))).
 user:message_property(debug(ods(test(error))), color(fg(red))).

webui.pl

@@ -44,7 +44,7 @@
 		   \table_rows(Sheet, SX,SY, EX,EY))), !.
 web_portray(cell(Sheet,X,Y)) -->
 	web_portray(cell_range(Sheet, X,Y, X,Y)).
-web_portray(table(_,_,Union)) -->
+web_portray(table(_Id,_Type,_DS,_Headers,Union)) -->
 	web_portray(Union).
 web_portray(List) -->
 	{ is_list(List), !,