The movtiation behind pyver, which is to creating a version control method which is very, very simple to track the history of files. The work-flow of pyver is something I do anyway when working on something simple that uses binary files, especially ones that are linked through the software that is required to read them (eg CAD designs). I will create multiple directories, eg v1, v2 and copy the entire contents of my in-progress files in it that I want to preserve the history. The issue with this is that when file names are changed, it can break links to dependent files, so keeping a file identical is important for tracking files. It was also a goal to use this tool over a network drive where multiple users can archive files as they wish There are a couple of other excellent options, but have a steep learning curve and some do not perform well with large binary files. Also, most version control systems are designed for code or text files, not large binanry files, like word documents or cad files. If you need more functionality, I'd recommend checking out these programs in order os simplicity. boar VCS, mercurial SCM, git with the large file extension
The basic operation of this program goes like this
1) create some files
2) run pyver and make a copy of your files
3) make as many revisions as you want
4) if you want to revert back to an older revision, grab files out of the dated .archive directories and put them in your current directory
Criteria for design
- Simplicity is your friend, obscurity is your enemy
- It is unacceptable to rename files to track versions. This accomplishes this by putting the files in different directories and keep your current file with the desired, persistent filename
- Can be used over a network drive. Does not require a central server.
- Be OS agnostic (written in Python)
- know when the changes were made, with a comment and who did it
To get started, lets do a quick demo in windows. Brace yourself, you are going to have to use the command line. In your current directory with the pyver.py file, open a command line with
shift+right_click and in the menu select
Open command window here. Another way to get a command window in your current directory is to type
alt+D and type
Done. That's it. We have created a pyver repository and made a backup of all the files in our current directory in an archive naming year-month-day-hour-minute-second. Anytime we need to make another backup, just repeat the last step.
OK, lets add a bit more work upfront to save us some typing for years to come. I would recommend having a windows folder that is added to the user PATH system variable where you can keep test or production scripts that you want to access from the command line. For example, I have
C:\Users\Neal\bin that is in the system PATH variable which I can dump python scripts with the *.bat file. Now, all we have to type is
pyver to get the same result as
To add this path variable, type
2) search "variables" and select "Edit Enviornmental variables for your account"
3) Add PATH variable or append with the path to the folder with your scripts
There are a few built in commands to interrogate the repository. The first is log, which shows the contents of a text file that keeps track of the commits that are made
pyver log Neal , 20160125210812 , 7kb , , pyver.py|README.md|pyver.bat Neal , 20160125210814 , 7kb , , pyver.py|README.md|pyver.bat Neal , 20160125210817 , 7kb , , pyver.py|README.md|pyver.bat Neal , 20160125210854 , 7kb , making a comment for this commit , pyver.py|README.md|pyver.bat Neal , 20160125210912 , 4kb , only adding one file , pyver.py
The next is tree, which shows all the files in the repo.
pyver tree ./.archive | ./20160125210812 | | pyver.py | | README.md | | pyver.bat | pyver.log | ./20160125210814 | | pyver.py | | README.md | | pyver.bat | ./20160125210817 | | pyver.py | | README.md | | pyver.bat | ./20160125210854 | | pyver.py | | README.md | | pyver.bat | ./20160125210912 | | pyver.py pyver.py README.md pyver.bat
We also have some flags we can use to customize our commits. This commit adds only the files file1.txt and file2.docx and adds a comment. Note-none of the flag inputs can have spaces, so files are seperated by
| and comments are written without spaces using
pyver -f "file1.txt|file2.docx" -c "making a comment for this commit"
We can also use wildcards to add groups of files rather than listing each individual file.
pyver -f "*.txt|*.docx" -c "I added all the txt and docx files to my archive"
Also, by default, pyver will skip all files and directories that start with . and ~. These files are typically temporary, backup, or configuration files.
I used pyinstaller to build my own windows executable
pyinstaller pyver.py --onefile
There you have it. The worlds easiest version control system that multiple users can use over a network, retrieving your old files is trivial, and now you don't have to change your file names to all the silly things people come up with.
You can find the source code here. http://github.com/nagordon/pyver
This was a fun project to learn more about file manipulation with python and in the end, I now have a handy tool!