How to pass bytes (base64) instead of string (utf-8) to streamGenerateContent?

[geotolky]

I would like to use the streamGenerateContent method to pass a file of any type (image, pdf, etc.) and then have the Gemini model answer a question about the image. The file would be local and not stored on CloudStorage.

Currently, in my Python notebook, I am:

1. reading in the contents of the file,
2. encoding them to base64 format (which looks like b“<string>” in Python), and
3. decoding to utf-8 format (“<string>” in Python).

I am then storing this in a JSON dictionary which I am passing to the model via an HTTP put request. This approach works fine. However, if I did want to send the data in base64 (b“<string>”) and essentially skip step 3 above, how can I do this?

Looking at the part of the above documentation which talks about the blob containing the data to be passed to the model, it says: “If possible send as text rather than raw bytes.” Here is a code example to illustrate what I mean:

import base64
import requests

with open(filename, 'rb') as f:
    file = base64.b64encode(f.read()).decode('utf-8') # HOW CAN I SKIP THE DECODING STEP?

url = ... # LINK TO GEMINI EXPERIMENTAL MODEL
headers = ... # BEARER TOKEN FOR AUTHORIZATION
data = { ...
          "inlineData": {
            "mimeType": "image/png", # OR "application/pdf" OR ANY OTHER FILE TYPE
            "data": string # HOW CAN I KEEP THIS AS A BASE64 OBJECT INSTEAD OF A STRING?
          },
         ...
        }

requests.put(url=url, json=data, headers=headers)

If the data were in bytes, this code would fail because JSON would not be able to serialize it. Alternatively, I tried using the data parameter in requests.put (data=json.dumps(data) instead of json=data), but I think this would also convert everything to a string, which wouldn’t work because I would like my input data to be in bytes format. Another possibility that I’ve seen is to use mimeType=“application/octet-stream”, but that’s not listed as a supported type in the documentation above.

Should I be using something other than JSON to make this request? Is what I’m describing even possible in Python? Any advice on how to make this work would be appreciated.

[Qasim_Taleb]

Hello
I am not experiencing in codes.but I will try:
you can use this:
import panda as pd
data_dict={"name":["name","name","name"},"age":[number,number,number],"city":["city
name","city name","city name"]}
df=pd.Data Frame(data_dict)
print(df)