How to calculate SHA from a string with the Github api in python?

Question

I want to update a file with the Github API and commit it into a branch. I have troubles creating the commit. The SHA does not match the expected one.

{
    'documentation_url': 'https://developer.github.com/enterprise/2.7/v3/repos/contents/',
    'message': 'pom.xml does not match de42fdd980f9b8067a2af982de46b8d5547e4597'
}

I do the following:

import hashlib
myfile = "new content of my README"
resulting_file = base64.b64encode(bytes(myfile, "utf-8"))
file_as_str = str(resulting_file.decode('utf-8'))
sha = hashlib.sha1(file_as_str.encode('utf-8')).hexdigest()

url = 'https://someurl.com/someproject/contents/README.md?access_token=' + access_token
data = '{"message": "bla bla", "content": "'+file_as_str+'", "sha": "'+sha+'", "branch": "'+branch+'"}'
response = requests.put(url, data=data)

I would not like to use a lib to do this to better understand what is happening. probably the SHA is not generated properly, but I cannot identify why. Could someone help?

Answer 1

You don't need to calculate the SHA for the new file. Instead you must supply the SHA of the file that is being replaced. You can obtain that by performing a get contents on the file using requests.get():

url = 'https://api.github.com/repos/someowner/someproject/contents/pom.xml'
r = requests.get(url)
sha = r.json()['sha']

Then use the value of sha in the PUT request to update the file:

with open('myfile', 'rb') as f:
    content = str(base64.b64encode(f.read()), encoding='utf8')
    data = {'message': 'bla bla', 'content': content, 'sha': sha, 'branch': branch}
    r = requests.put(url, json=data)

Answer 2

GitHub calculates hashes as followes:

sha1("blob " + filesize + "\0" + data)

so use this code:

with open(filepath, 'rb') as file_for_hash:
    data = file_for_hash.read()
    filesize = len(data)
sha = hashlib.sha1("blob " + filesize + "\0" + data).hexdigest()