threads / discuss / 16028

clean/smudge filters for pdf files

Subject: clean/smudge filters for pdf files

## tl;dr

5 messages between Oct 23, 2008 and Oct 24, 2008.

replies: 4people: 3as markdown or json

Leo Razoumov· Oct 23, 2008, 19:44 UTC · lore

I am trying to improve storage efficiency for PDF files in a git repo. Following earlier discussions in this list I am trying to set up proper clean/smudge filters. What follows is my current setup

# in ~/.gitconfig
[filter "pdf"]
	clean  = "pdftk - output - uncompress"
	smudge = "pdftk - output - compress"

# in .gitattributes *.pdf filter=pdf

Unfortunately, it seems as though that pdftk uncompress followed by pdftk compress do not leave the file invariant. I tried several uncompress+compress iterations and the file still keep changing (the size though stays the same). Is there any other alternative way to store PDF files in git repo more efficiently? Any alternative to pdftk on Linux?

--Leo--

Pierre Habouzit· Oct 23, 2008, 21:32 UTC · re: Leo Razoumov · lore

Re: clean/smudge filters for pdf files

On Thu, Oct 23, 2008 at 07:44:39PM +0000, Leo Razoumov wrote:

Show 19 quoted lines

> I am trying to improve storage efficiency for PDF files in a git repo.
> Following earlier discussions in this list I am trying to set up
> proper clean/smudge filters. What follows is my current setup
> 
> # in ~/.gitconfig
> [filter "pdf"]
> 	clean  = "pdftk - output - uncompress"
> 	smudge = "pdftk - output - compress"
> 
> # in .gitattributes
> *.pdf filter=pdf
> 
> Unfortunately, it seems as though that pdftk uncompress followed by
> pdftk compress do not leave the file invariant. I tried several
> uncompress+compress iterations and the file still keep changing (the
> size though stays the same).
> Is there any other alternative way to store PDF files in git repo more
> efficiently?
> Any alternative to pdftk on Linux?

actually it uses some kind of zlib algorithm so that's pretty normal you don't have the same result with a packer. Maybe one could write a tool like pristine-tar for that purpose.

-- 
·O·  Pierre Habouzit
··O                                                madcoder@debian.org
OOO                                                http://www.madism.org

Leo Razoumov· Oct 24, 2008, 01:40 UTC · re: Pierre Habouzit · lore

Re: clean/smudge filters for pdf files

On 10/23/08, Pierre Habouzit <madcoder@debian.org> wrote:

Show 26 quoted lines

> On Thu, Oct 23, 2008 at 07:44:39PM +0000, Leo Razoumov wrote:
>  > I am trying to improve storage efficiency for PDF files in a git repo.
>  > Following earlier discussions in this list I am trying to set up
>  > proper clean/smudge filters. What follows is my current setup
>  >
>  > # in ~/.gitconfig
>  > [filter "pdf"]
>  >       clean  = "pdftk - output - uncompress"
>  >       smudge = "pdftk - output - compress"
>  >
>  > # in .gitattributes
>  > *.pdf filter=pdf
>  >
>  > Unfortunately, it seems as though that pdftk uncompress followed by
>  > pdftk compress do not leave the file invariant. I tried several
>  > uncompress+compress iterations and the file still keep changing (the
>  > size though stays the same).
>  > Is there any other alternative way to store PDF files in git repo more
>  > efficiently?
>  > Any alternative to pdftk on Linux?
>
>
> actually it uses some kind of zlib algorithm so that's pretty normal you
>  don't have the same result with a packer. Maybe one could write a tool
>  like pristine-tar for that purpose.
>

With zlib you get the same deterministic result as long as you use the same zlib packer and unpacker. With pdftk compress/uncompress seem not to form a bijection pair. This issue was briefly discussed on this list back in April 2008 but no resolution emerged.

--Leo--

Michael J Gruber· Oct 24, 2008, 08:10 UTC · re: Leo Razoumov · lore

Re: clean/smudge filters for pdf files

Leo Razoumov venit, vidit, dixit 24.10.2008 03:40:

Show 32 quoted lines

> On 10/23/08, Pierre Habouzit <madcoder@debian.org> wrote:
>> On Thu, Oct 23, 2008 at 07:44:39PM +0000, Leo Razoumov wrote:
>>  > I am trying to improve storage efficiency for PDF files in a git repo.
>>  > Following earlier discussions in this list I am trying to set up
>>  > proper clean/smudge filters. What follows is my current setup
>>  >
>>  > # in ~/.gitconfig
>>  > [filter "pdf"]
>>  >       clean  = "pdftk - output - uncompress"
>>  >       smudge = "pdftk - output - compress"
>>  >
>>  > # in .gitattributes
>>  > *.pdf filter=pdf
>>  >
>>  > Unfortunately, it seems as though that pdftk uncompress followed by
>>  > pdftk compress do not leave the file invariant. I tried several
>>  > uncompress+compress iterations and the file still keep changing (the
>>  > size though stays the same).
>>  > Is there any other alternative way to store PDF files in git repo more
>>  > efficiently?
>>  > Any alternative to pdftk on Linux?
>>
>>
>> actually it uses some kind of zlib algorithm so that's pretty normal you
>>  don't have the same result with a packer. Maybe one could write a tool
>>  like pristine-tar for that purpose.
>>
> 
> With zlib you get the same deterministic result as long as you use the
> same zlib packer and unpacker. With pdftk compress/uncompress seem not
> to form a bijection pair. This issue was briefly discussed on this
> list back in April 2008 but no resolution emerged.

For a different file format I use the pair "gzip -c, gunzip -c" without any problems, so zlib is not a problem. I do see the effect that checkouts on different machines may have different compressed files (same gzip version), but this is a non-issue.

Your experience with pdftk confirms mine. It shuffles things around becauses it parses the files into objects and then writes them out again in possibly different order. This is no problem for pdf because it uses "pointers" (it's a bijection up to reordering), but it's a weird design, and complicates things for us.

I'm still looking for something viable, I'll let list know when I've found something...

Michael

Michael J Gruber· Oct 24, 2008, 08:44 UTC · re: Leo Razoumov · lore

Re: clean/smudge filters for pdf files

Little addition to my previous reply:

Multivalent apparently almost get's there. After 2 iterations most of the uncompressed file is stable, except for some binary blob at the end. Alas, it's Java and not even completely open source.

Michael