Use the same table number in the source id and the saved files

[SashaMIT]

parse_all_tables writes the spreadsheet and the image as table{ctr}, then increments ctr, then stores table{ctr} in the document source. The first table is saved as table1 and reported as table2. The second is saved as table2 and reported as table3.

The same increment sits in both community copies of the parser. The source id is now taken before the counter moves, so it matches the files just written.

The new test stubs the PDF and model imports and checks both copies. On tip the first source is report-page0-table2 while the file is table1-page0.xlsx. After the change both sources are table1 and table2.