find_tables on landscape page generates reversed text

[dsanchezseco-ibm]

Describe the bug (mandatory)

When performing a find_tables over a table in a landscape page the extracted text, either with extract or with to_pandas generates the text and the lists to be reversed

To Reproduce (mandatory)

Execute this snippet over a table in a landscape page

def extract_tables(pdf_document, start, end):
    tables = []
    for page in range(start, end + 1):
        local_tables = pdf_document[page].find_tables()
        for table in local_tables:
            tables.append(table.extract())
    print(tables)

it will generate something like

[[['aicnerefeR\n)CRM(', '100', '200', '300', '400', '500', '600', '700', '800', '900'], ['abeurp\n,)sotseuper(\ned\nsopiuqe\nselbimusnoc\n,]sodireuqer\ny\nsatneimarreH[', 'acirtémomanid acirtémomanid\nevalL evalL\n]1DL[ ]2DL[', 'ó\notcatnoc\naselgnI\nnis evalL\nresál litátrop\n]IL[\nortemómreT ó\nrapomreT ajiF\nevalL\n]LT[ ]PT[ ]FL[', 'nóicarbiv\ned\nrodideM\n]VM[', 'esargne\n)K2K\nalotsiP\nasarg(\n]EP[', 'serotom\narap\nlitátrop\naicnetop\nortemóucavonaM\nortemónaM\ned\nrodideM\nsocisáfirt\n]50MM[ ]50VM[\n]WM[', 'aselgnI\n)rodaicnatsid )rotom\n)abmob )otneimalpoca\nevalL roiretxe )ocinácem )otelpmoc )otelpmoc otneimador\nroiretni\n)etsagsed\nocitsálp otneimador\n]IL[ó latigid\nnellA ortemórciM ortemórciM )roslupmi )rotcelfed olliuqsac\nollitraM ajiF erbilaC ereic ollina abmob )satnuj socat otneimalpoca( 3C/6136\nevalL evalL\n0222( 0452( 1.0024( 0051( ed 1.0642( 1103x2( ogeuj( ogeuj(\n]PM[ ]FL[ ]AL[ ]IM[ ]EM[ ]DC[ eje( x2(', 'aselgnI\nevalL\n)ocinácem\n]IL[\nlatigid\nó nellA\najiF erreic )satnuj\nerbilaC\nevalL evalL\n1.0024(\nogeuj(\n]FL[ ]AL[ ]DC[', 'aselgnI\n)rodaicnatsid\n)abmob\nevalL\nocitsálp otneimador\n]IL[\nlatigid\nolliuqsac\nó nellA\najiF )satnuj\nerbilaC ollitraM\nevalL evalL 1103\n1.0642(\nogeuj(\n]FL[ ]AL[ ]DC[ ]PM[ x\n2(', 'aselgnI\n)rotom\nevalL\notneimador\n]IL[\nlatigid\nó nellA\najiF\nerbilaC 3C/6136\nevalL evalL\n]FL[ ]AL[ ]DC[ x\n2('], ['aírogetaC lanosrep', 'ocináceM', 'ocináceM', 'ocináceM', 'ocináceM', '&\natsicirtcelE ocináceM', 'ocináceM', 'ocináceM', 'ocináceM', 'ocináceM'], ['odirrucsnart\nopmeiT )saroh(', '5,0', '80,0', '1', '1,0', '1', '4', '2', '3', '1'], ['daditnaC erbmoh', '1', '1', '1', '1', '2', '2', '2', '2', '1'], [')saroh(\nlatot\nopmeiT odamitse', '5,0', '80,0', '1', '1,0', '2', '8', '4', '6', '1'], ['dadicidoireP', 'lausneM', 'lartsemirT', 'lartsemirT', 'lartsemirT', 'launA', 'launaiB', 'ovitcerroC\n)aguf\nis(', ')arutor\novitcerroC\nyah\nis(', ')arutor\novitcerroC\nyah\nis('], ['le', 'odr', 'odr', 'odr', 'odr', '', '', '', '', '']]]

For problems when building or installing PyMuPDF, give the full output of the build/install command so that, for example, all pip/compiler/linker errors/warnings can be seen.

Expected behavior (optional)

expected for this case is

[[['009', '008', '007', '006', '005', '004', '003', '002', '001', '(MRC)\nReferencia'], ['(2\nx [CD] [LA] [LF]\nLlave Llave\n6316/C3 Calibre\nFija\nAllen ó\ndigital\n[LI]\nrodamiento\nLlave\nmotor)\nInglesa', '(2\nx [MP] [CD] [LA] [LF]\n(juego\n(2460.1\n3011 Llave Llave\nMartillo Calibre\njuntas) Fija\nAllen ó\ncasquillo\ndigital\n[LI]\nrodamiento plástico\nLlave\nbomba)\ndistanciador)\nInglesa', '[CD] [LA] [LF]\n(juego\n(4200.1\nLlave Llave\nCalibre\njuntas) cierre Fija\nAllen ó\ndigital\n[LI]\nmecánico)\nLlave\nInglesa', '(2x (eje [CD] [ME] [MI] [LA] [LF] [MP]\n(juego (juego (2x3011 (2460.1 de (1500 (4200.1 (2540 (2220\nLlave Llave\n6316/C3 (acoplamiento tacos juntas) bomba anillo ciere Calibre Fija Martillo\ncasquillo deflector) impulsor) Micrómetro Micrómetro Allen\ndigital ó[LI]\nrodamiento plástico\ndesgaste)\ninterior\nrodamiento completo) completo) mecánico) exterior Llave\nacoplamiento) bomba)\nmotor) distanciador)\nInglesa', '[MW]\n[MV05] [MM05]\ntrifásicos\nMedidor\nde\nManómetro\nManovacuómetro\npotencia\nportátil\npara\nmotores', '[PE]\n(grasa\nPistola\nK2K)\nengrase', '[MV]\nMedidor\nde\nvibración', '[LF] [TP] [TL]\nLlave\nFija Termopar\nó Termómetro\n[LI]\nportátil láser\nLlave sin\nInglesa\ncontacto\nó', '[LD2] [LD1]\nLlave Llave\ndinamométrica dinamométrica', '[Herramientas\ny\nrequeridos],\nconsumibles\nequipos\nde\n(repuestos),\nprueba'], ['Mecánico', 'Mecánico', 'Mecánico', 'Mecánico', 'Mecánico Electricista\n&', 'Mecánico', 'Mecánico', 'Mecánico', 'Mecánico', 'personal Categoría'], ['1', '3', '2', '4', '1', '0,1', '1', '0,08', '0,5', '(horas) Tiempo\ntranscurrido'], ['1', '2', '2', '2', '2', '1', '1', '1', '1', 'hombre Cantidad'], ['1', '6', '4', '8', '2', '0,1', '1', '0,08', '0,5', 'estimado Tiempo\ntotal\n(horas)'], ['(si\nhay\nCorrectivo\nrotura)', '(si\nhay\nCorrectivo\nrotura)', '(si\nfuga)\nCorrectivo', 'Bianual', 'Anual', 'Trimestral', 'Trimestral', 'Trimestral', 'Mensual', 'Periodicidad'], ['', '', '', '', '', 'rdo', 'rdo', 'rdo', 'rdo', 'el']]]

Your configuration (mandatory)

3.11.6 (main, Oct  2 2023, 13:45:54) [Clang 15.0.0 (clang-1500.0.40.1)]
 darwin

PyMuPDF 1.23.6: Python bindings for the MuPDF 1.23.5 library.
Version date: 2023-11-06 00:00:01.
Built for Python 3.11 on darwin (64-bit).

PyMuPDF installed with pip

Additional context (optional)

Add any other context about the problem here.

[JorjMcKie]

I would like to reproduce this. Can you please provide an example page?
Thanks.

[JorjMcKie]

Wait for official publication before closing this.

[dsanchezseco-ibm]

Cool, thank you! I'll keep an eye out. Sorry for not providing an example, this was the work account and I was on weekend mode

[JorjMcKie]

Cool, thank you! I'll keep an eye out. Sorry for not providing an example, this was the work account and I was on weekend mode

No problem. The necessary changes are confined to one file of the package, table.py.
Don't know about your update privileges on your work account, but there is the hot fix option to just replace that file in your installation folder and confirm the problem is indeed solved.

[dsanchezseco-ibm]

I'll check it out and let you know

[dsanchezseco-ibm]

https://github.com/pymupdf/PyMuPDF/pull/2818/files#diff-93d282fd72c8edf2245fbab4a988f37011fea2c8db71a8b7441116e867ddeabaR1168

shouldn't it be self.cells?

installed the current main and it is complaining on that line about the cells var

[JorjMcKie]

Absolutely! Thank you!

[dsanchezseco-ibm]

ok, the table extraction does work now correctly(trying the hot-fix) but there is another typo, you are assigning r instead of rect

[JorjMcKie]

incredible - it's Monday morning 😒
Thanks again!

[dsanchezseco-ibm]

All good now! Thanks and enjoy the coffee!

[julian-smith-artifex-com]

Fixed in 1.23.7.