wrong return

[Larbo53]

Hello,

I extract a zone with the command :
invoice=doc.pages[i].crop((row[2],row[3],row[4],row[5]),relative = True).extract_text()

the returned value is : 9161F42A87C
while the display in the pdf is : 91614287.
How can I find the right value? Is it a problem with the fonts or the pdf editor?

Thanks for your feedback.

Best regards.
Valhrona_Facture n°91614287.pdf

[jsvine]

What is row here? Or, more generally, can you share a full code snippet that reproduces your issue?

[jsvine]

How to knows the pdfplumber’s version?

Run pip freeze in the environment where you're using pdfplumber, and see what the line for pdfplumber=... says.

2 - I don't understand your second remark :

I was suggesting that you use Markdown code blocks, to make the code in your comments easier for others to read: https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/creating-and-highlighting-code-blocks

[Larbo53]

Hi,

with another invoice and the program, it's the same problem, with the same environment :
Result :
{'CN00° 1fa6c0t3u0re04': [1], 'CN00° 1fa5c9t7u5re91': [2], 'CN00° 1fa5c9t3u3re47': [3], 'CN00° 1fa5c8t9u6re26': [4], 'CN00° 1fa5c8t5u7re32': [5], 'CN00° 1fa5c8t0u9re95': [6], 'CN00° 1fa5c7t2u9re65': [7]}
rec : (399.93, 41.83, 461.03, 58.03)
pdf join

What's wrong?

Sincerely
Deroche_Factures C109046.pdf

[jsvine]

Thanks! I can reproduce your issue with this one, and now I see what's happening: The box you're drawing (displayed below in green) overlaps slightly with the bounding boxes of the characters (displayed in red) in the line above the one you want:

Because .crop(...) retains any object that overlap with the bounding box provided, those unwanted characters are still in the cropped version. There are a few different ways you could resolve this:

• Use .within_bbox(...), which only retains objects fulling within the bounding box provided.
• Or: Change your bounding box to be slightly smaller, so that it doesn't overlap with the characters above.
• Or: Keep your current approach, but filter out any objects with a height less than, say 3, which will remove the objects that only overlapped a little bit.

[Larbo53]

hi,

it's OK with .within_bbox(...)

Thanks a lot.

[Larbo53]

Good morning,

I'm still using the same process, but with a pdf file that contains several invoices and blank pages between some of them.
If the processing finds a blank page, it stops. What is the solution?

My code :
for i in range(0,len(doc.pages)):

invoice=doc.pages[i].within_bbox(rec,relative = True).extract_text()
if invoice[-2] not in ('0','1','2','3','4','5','6','7','8','9') and len(invoice)!=0:
    continue
else:

Pages_Factures[doc.pages[i]]=invoice

with rec=(399.93, 41.83, 461.03, 58.03)

Thanks a lot
deroche C108308.pdf

[Larbo53]

Hi, With pip3 list : the pdfplumber’s version is 0.6.0 Yves Larbodiere EVARD 3, rue des Courtes Terres 95220 HERBLAY portable : 07 81 08 41 00 mail : ***@***.***

…

Le 3 févr. 2022 à 23:31, Jeremy Singer-Vine ***@***.***> a écrit : How to knows the pdfplumber’s version? Run pip freeze the environment where you're using pdfplumber, and see what the line for pdfplumber=... says. 2 - I don't understand your second remark : I was suggesting that you use Markdown code blocks, to make the code in your comments easier for others to read: https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/creating-and-highlighting-code-blocks <https://docs.github.com/en/get-started/writing-on-github/working-with-advanced-formatting/creating-and-highlighting-code-blocks> — Reply to this email directly, view it on GitHub <#596 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AU2RHDAR2PIMFFYY47DO6YLUZL64NANCNFSM5NPH4YJQ>. Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you authored the thread.

[Larbo53]

Hi,

today, with another invoice with the same environnement and the same program, I have the same problem 👍

metadata : {'Producer': 'PDFCreator Free 3.4.1', 'CreationDate': "D:20210701180908+02'00'", 'ModDate': "D:20210701180908+02'00'", 'Title': 'Facture C001603004 - Etat', 'Author': 'jmistretta', 'Subject': '', 'Keywords': '', 'Creator': 'PDFCreator Free 3.4.1'}

bbox : (399.93, 41.83, 461.03, 58.03)
return : {'CN00° 1fa6c0t3u0re04': [1], 'CN00° 1fa5c9t7u5re91': [2], 'CN00° 1fa5c9t3u3re47': [3], 'CN00° 1fa5c8t9u6re26': [4], 'CN00° 1fa5c8t5u7re32': [5], 'CN00° 1fa5c8t0u9re95': [6], 'CN00° 1fa5c7t2u9re65': [7]}
pdf joint

what's wrong?

Sincerely
Deroche_Factures C109046.pdf

[Larbo53]

Hi,

I can't find my request of yesterday about the management of white pages in a pdf(attached) which groups several invoices.
here is my code :
for i in range(0,len(doc.pages)):

invoice=doc.pages[i].within_bbox(rec,relative = True).extract_text()
if invoice[-2] not in ('0','1','2','3','4','5','6','7','8','9') and len(invoice)!=0:
    continue
else:
    Pages_Invoices[doc.pages[i]]=invoice

with rec=(399.93, 41.83, 461.03, 58.03)

The program is blocked by the first blank page.
What is the solution?

Thank you.

Sincerely

Translated with www.DeepL.com/Translator (free version)
deroche C108308.pdf

[tegancp]

I would think you're getting an IndexError for a blank page, since your invoice variable in that case does not have an element at index -2 to access. Try switching the order of your conditional, so that you're checking first that invoice has elements before accessing them.

[Larbo53]

hi,

The result is same.
What value returns an area or a blank page? By checking this value, the program could then read all the pages in the file.

Thanks for your feedback.
Best regards.

[Larbo53]

with 👍

for i in range(0,len(doc.pages)):#vérifier que c'est un num conforme c'est à dire qu'il contient au moins 2 chiffres

invoice=doc.pages[i].within_bbox(rec,relative = True).extract_text()
if invoice=='':
    continue

elif invoice[-2] in ('0','1','2','3','4','5','6','7','8','9'):
    Pages_Factures[doc.pages[i]]=invoice

it's OK.

Thanks a lot.

Can you close this discussion?