wish for open method

[CBusse]

May I suggest some more advanced interface into a SevenZipFile with an open method like it is provided in the zipfile module? Supporting read and write mode at minimum.

I am desperately looking into a way to use py7zr to generate bigger data sets that should not be kept in a simple data block/array, but need to be streamed into the archive (not from file system, either from another archive, or for example from a database).

For this I'd like to have some kind of a file handle into the archive where I could do writes of flexible sizes (just like a normal file handle) and data types.

While there is the writestr method, this method just does not give me the flexibility I would like to have (it is just not sensible to keep a structure that could go into multiple gigabyte sizes in memory all the time to supply as parameter here).

Do you see an easy path to implement this?

[miurahr]

Proposal and test implementation is welcome. Because I do not use memory interface by myself, I do not have a preference for it.
Also because of a nature of 7zip format, it is not possible to implement a file-system like interface that zipfile module provide.
7zip is a solid compression format. when you need last several bytes from a gigabytes of archive, you need to extract all the previous compressed data and then you get a byte.

[avm19]

7zip is a solid compression format. when you need last several bytes from a gigabytes of archive, you need to extract all the previous compressed data and then you get a byte

@miurahr Does this mean that it is impossible to read a specific file from a 7z archive in a streaming mode?

I see that the latest (1.0.0-rc2) docs show an example of how to pass a custom file-like object factory to py7zr so that it writes into it except of writing to the default disk filesystem.

I wonder if it is possible to get a file-like object from py7zr for reading? Again, just like in this Python stdlib ZipFile example, where a file-like object supports read(size=-1) method? In my use case I extract the contents of archive.7z on disk, open each resulting file and pass them as streams. I want to avoid extracting files on disk and read directly from the compressed archive, even if this incurs some Python overhead.