How do I “create”/“assign” a logging handler for Google Cloud Pubsub?

Question

Development from the previous thread found that the assumptions when asking the question were off-topic (subprocess was actually not causing the problems), so I\'m making a more

Answer 1

First, if I understand correctly by what you are showing in your output, you are using a Pub/Sub notification to send a message whenever you make changes to a Cloud Storage object. This information could be helpful.

Now, message.data["name"] is not going to work because message.data is a BYTES object. Thus, can't be indexed as a dict.

To treat it as a dict, you first have to decode it as base64 (import base64). After that, what you are left is a string which looks like JSON format. You then use json.load() (don't forget to import json) to transform this string into a dict. Now you can index the message.

The code for this will be:

print("This is before variables")
dirpath = "/subfolder1/"
print(dirpath)

#Transform the bytes object into a string by decoding it
namepath = base64.b64decode(message.data).decode('utf-8')

#Transform the json formated string into a dict
namepath = json.loads(namepath)

print(namepath["name"])
fullpath = dirpath + namepath["name"]
print(fullpath)
print("this is after variables")

Now, if your intent is to read the attributes only, they are properly defined at the top like:

    if message.attributes:
        print('Attributes:')
        for key in message.attributes:
            value = message.attributes.get(key)
            print('{}: {}'.format(key, value))

So, you could use:

    print("this is before variables")
    dirpath = "~/subfolder1/"
    print(dirpath)
    namepath = message.attributes["objectId"]
    print(namepath)
    fullpath = dirpath + namepath
    print(fullpath)
    print("this is after variables")

Keep in mind that for this particular case, "objectId" is the name of the file because it's the attribute that the notification from Pub/Sub for Cloud Storage uses. If you pretend to send custom messages, change "objectId" to your desired attribute name.

Answer 2

As Nahuel and tripleee explained, the problem is with the messages being BYTES instead of strings. However, their code didn't exactly work, and still threw out errors, and I have no idea why. By cross-referencing with Google's sample code for the pubsub appengine website, and a few more hours of trial and error, I have found the following code to be working. Might be inelegant and/or have bad practices, in that case please edit it and make it more robust.

#Continues from after message.ack(), above code remains unchanged
#except needing to <import json>

    #this makes a message.data a true python dict with strings.
    payload = json.loads(message.data.decode('utf-8')) 

    #this finds the value of the dict with key "name"
    namepath = payload["name"]

    #this is just a static string to pre-pend to the file path
    dirpath = "/home/[redacted]/"

    #combine them into a single functioning path
    fullpath = dirpath + namepath

    #currently type 'unicode', so convert them to type 'str'
    fullpath = fullpath.encode("utf-8")

And at the end we will have a fullpath that is purely type 'str' to be used by later functions/commands.