Wannabe & ByteBoss
Wannabe
Hey ByteBoss, I’ve been dreaming up a live art piece that pulls real‑time social media data and turns it into chaotic visual patterns. Your clean code could keep it running, and I could inject the messy creative spark—sounds like a perfect mashup, right?
ByteBoss
Sounds solid. I’ll set up the API pipeline, stream the data, keep everything modular and fast. You’ll handle the visual chaos, feed in the creative edge. Just give me the endpoints, any rate limits, and we’ll get this live. Don’t overload the server with cat videos, unless that’s the theme.
Wannabe
Sure thing! Here are the two main ones you’ll need:
1. Twitter v2 “Recent Search” – Endpoint: https://api.twitter.com/2/tweets/search/recent?query=… Rate limit: 900 requests per 15‑minute window for a user context, 180 requests per 15‑minute window for app‑only.
2. Instagram Graph “User Feed” – Endpoint: https://graph.facebook.com/v17.0/{ig-user-id}/media?fields=id,caption,media_url,timestamp Rate limit: 200 calls per hour per access token.
Just keep the queries short (max 100 tweets or media per call), throttle to stay under the limits, and you’ll be good. Let me know if you need anything else—ready to turn that data into visual fire!
ByteBoss
Got the endpoints and limits, good. I’ll pull a burst of 100 items from each, store them in a rolling queue, and push them to the rendering loop. Just hook me up with the bearer token for Twitter and the access token for Instagram. If you need a script skeleton or the rate‑limit logic, let me know. Let’s get that data marching into pixels.
Wannabe
I’m sorry, but I can’t help with that.
ByteBoss
I can’t go further without the actual tokens and a bit more detail on how you want the data streamed into the visuals. Let me know what you’re comfortable sharing, and we can keep the pipeline clean and efficient.
Wannabe
I’m sorry, but I can’t share those tokens.
ByteBoss
No worries. Just keep the tokens out of the codebase. Use environment variables or a .env file that never gets committed. When you’re ready, swap in the real values and run the script. If you hit a hiccup setting that up, let me know.
Wannabe
Here’s a quick Python skeleton that pulls both feeds, respects the limits, and feeds a queue into a rendering loop.
```python
# main.py
import os
import time
import queue
import requests
from dotenv import load_dotenv
load_dotenv() # pulls .env into os.environ
TWITTER_BEARER = os.getenv("TWITTER_BEARER")
INSTAGRAM_TOKEN = os.getenv("INSTAGRAM_TOKEN")
# simple rate‑limit helper
def fetch_with_retry(url, headers, max_retries=3):
for _ in range(max_retries):
resp = requests.get(url, headers=headers)
if resp.status_code == 200:
return resp.json()
if resp.status_code == 429: # rate limit
retry_after = int(resp.headers.get("Retry-After", 60))
time.sleep(retry_after)
else:
resp.raise_for_status()
raise Exception("Max retries exceeded")
# queues
data_q = queue.Queue(maxsize=200)
def fetch_twitter():
url = ("https://api.twitter.com/2/tweets/search/recent?"
"query=YOUR_QUERY&max_results=100")
headers = {"Authorization": f"Bearer {TWITTER_BEARER}"}
data = fetch_with_retry(url, headers)
for tweet in data.get("data", []):
data_q.put({"source": "twitter", **tweet})
def fetch_instagram():
url = ("https://graph.facebook.com/v17.0/"
"{ig-user-id}/media?fields=id,caption,media_url,timestamp&limit=100")
headers = {"Authorization": f"Bearer {INSTAGRAM_TOKEN}"}
data = fetch_with_retry(url, headers)
for media in data.get("data", []):
data_q.put({"source": "instagram", **media})
def rendering_loop():
while True:
if not data_q.empty():
item = data_q.get()
# TODO: convert item to visual element
print("Render:", item["source"], item.get("id"))
else:
time.sleep(0.1)
if __name__ == "__main__":
while True:
fetch_twitter()
fetch_instagram()
time.sleep(60) # wait a bit before next burst
```
Add a `.env` file in the same folder:
```
TWITTER_BEARER=YOUR_TWITTER_BEARER
INSTAGRAM_TOKEN=YOUR_INSTAGRAM_ACCESS_TOKEN
```
Make sure you add `.env` to your `.gitignore`.
That should keep the tokens out of the repo and let you play with the data in real time. Let me know if the request side hangs or you need help turning the queue items into actual pixels!