Extracttext: error

[yanfei]

How to check whether the configuration of tesseract is correct in proxmox? At present, I can always find warning information in the log, but I can get information by manually executing tesseract. If I don't specify the language, it will be garbled. When I manually specify the language as simplified Chinese, it can be correctly recognized. How to configure the execution parameters of tesseract in PMG?

Dec 2 09:53:31 sh-pmg pmg-smtp-filter[195600]: WARNING: extracttext: error (1) from /usr/bin/tesseract: Premature end of JPEG file

test execute Command:
tesseract wechat.png output
tesseract wechat.png output -l chi_sim




wechat.png

 

[yanfei]

I found that the /etc/mail/spamassassin/v400.pre configuration file contains extracttext configuration information. I manually modified the tesseract default command to extracttext_external tesseract {OMP_THREAD_LIMIT=1} /usr/bin/tesseract -l chi_sim+eng -c page_separator= {} -, is it possible to do this? I have installed tesseract in both Simplified Chinese and Traditional Chinese.

 

[Stoiko Ivanov]

Dec 2 09:53:31 sh-pmg pmg-smtp-filter[195600]: WARNING: extracttext: error (1) from /usr/bin/tesseract: Premature end of JPEG file

on a hunch - how large is the email (go by the information from the journal/mail-log? and how large is the max spamsize you've set (GUI->Configuration->Spam Detector->Options) ? - ensure that max-spamsize is as large as your general mail-size (we'll improve this in one of the next versions as this seems to cause issues for more users)

 

[yanfei]

I used the default parameters of 262144, and it seems that the error message size does exceed these sizes.

on a hunch - how large is the email (go by the information from the journal/mail-log? and how large is the max spamsize you've set (GUI->Configuration->Spam Detector->Options) ? - ensure that max-spamsize is as large as your general mail-size (we'll improve this in one of the next versions as this seems to cause issues for more users)

 

[Stoiko Ivanov]

I used the default parameters of 262144, and it seems that the error message size does exceed these sizes.

set it to your Message Size (GUI->Configuration->Mail Proxy->Options) - and see if it works then.

 

[yanfei]

set it to your Message Size (GUI->Configuration->Mail Proxy->Options) - and see if it works then.

OK, I'll give it a try.
I modified the default configuration file /etc/mail/spamassassin/v400.pre to
extracttext_external tesseract {OMP_THREAD_LIMIT=1} /usr/bin/tesseract -l chi_sim+eng -c page_separator= {} -
Is it possible? I hope to recognize Chinese.

 

[Stoiko Ivanov]

please try it first only with the size setting - if this does not work - adapt your modification and see if this helps