juanirm / pst-converter

A docker container with libpst will convert all pst files from an input folder to mbox format on an output folder

PST Converter

This small project aims to provide a free and open source way to convert .pst files (personal storage table) to MBOX files, structured in a way that Mozilla Thunderbird can import.

How it works

A docker container is run with the libpst library inside. The container invokes the converter.sh Bash script. The Bash script executes the readpst utility which converts all .pst files inside the in directory, and outputs them to the out directory. find commands are then run to restructure the directories so Thunderbird can import them.

Usage

1. Put the .pst files to be converted into the in folder
2. Build the docker image: docker-compose build
3. Start conversion: docker-compose up
4. The script will produce a single directory with an .sbd suffix (short for sub-directory). Example: 5db2f19c41144330b8c75bc1cb37d3cc.sbd

How to import into Thunderbird

You need to move the contents of the .sbd directory to a Thunderbird profile directory. To find this, go to Account Settings in Thunderbird, click Local Folders in the sidebar (or your actual email account if you want to import the messages to that). Under the Message Storage heading there will be a path - move the contents of the .sbd directory to that. Restart Thunderbird.

How to customize it

Conversion options can be set in the convert.sh file Input and output directories can be changed in docker-compose.yml. Keep in mind that the convert.sh script does some magic based on the output directory being out. Some later versions of this might make better usage of environment variables to 'fix' this 🙂

Credits

• Paul Colby for his article: http://colby.id.au/importing-pst-files-into-thunderbird-using-libpst/
• libpst team for their tool: https://www.five-ten-sg.com/libpst/

Resources

• readpst man page: https://www.five-ten-sg.com/libpst/rn01re01.html

pst-converter's Issues

benefit of readpst -u versus -r conversion

hello,
after playing more with this converter i find that the readpst -u option creates extra .size and .type files which are ignored by my Thunderbird (linux, v115.6.0).

Furtermore the .size and .type files make that the last step in the conversion always fails, child folders are never empty due to the generated .size and .type files. This gives more cluttered output.

Is there a reason to use -u versus the regular -r which is used on most examples on internet?

To compare results I have run the same .pst file with both -u and -r options and the mbox content is identical apart from
* - - - -boundary-LibPST-iamunique- [number] *
lines. But upon further investigation each run generates new numbers for these meta items.

Thunderbird does not seem to support .size and .type files:
http://kb.mozillazine.org/Profile_folder_-_Thunderbird#Files
I searched my profile folder with a handful of mail accounts and dozens of folders but no such files exist.

My proposal would be to switch to readpst with -r option.

How do I get converted files into Thunderbird?

The conversion was successful however I can't import the output files into Thunderbird. How do I do this? The files are HTML not MBOX like the README says.

Also, what are the .sbd directories? There is no mention of these in the README.

Example of the file types I get as output:

$ file out/5db2f19c41144330b8c75bc1cb37d3cc.sbd/Inbox
Inbox: HTML document, Non-ISO extended-ASCII text, with very long lines (961)

converting more than one pst

hello,
when I put more than one pst file in the in/ folder only the first gets processed.
Looking at the manpage of readpst https://www.five-ten-sg.com/libpst/rn01re01.html it seems it accepts one file as input.

So I simply updated convert.sh to loop all pst files in the in/ folder like this:

#!/bin/bash
for INDIRfile in $INDIR/*.pst
do
  readpst -u -o $OUTDIR $INDIRfile 
done
...