tfriedel

Keybase proof

I hereby claim:

• I am tfriedel on github.
• I am tfriedel (https://keybase.io/tfriedel) on keybase.
• I have a public key ASCVWUETl51ZIgGRxm6ZKXXWU5VSAhjQ4V0ocVJ84nKybgo

To claim this, I am signing this object:

tfriedel

!(function (e, t) {
    "object" == typeof exports && "undefined" != typeof module ? t(exports) : "function" == typeof define && define.amd ? define(["exports"], t) : t((e.Labelbox = {}));
})(this, function (exports) {
    "use strict";
    var __assign = function () {
            return (__assign =
                Object.assign ||
                function (e) {
                    for (var t, r = 1, n = arguments.length; r < n; r++) for (var i in (t = arguments[r])) Object.prototype.hasOwnProperty.call(t, i) && (e[i] = t[i]);
                    return e;

tfriedel

Vision API Exporter

You:

I need to analyse bulk images (50-100) and export results as csv. I want to use the Chatgpt Vision api ( https://platform.openai.com/docs/guides/vision ).

Modify this code to export results:

import base64
import requests

tfriedel

"""Bulk-download every run from spire-codex.com.

Lists all runs from ``/api/runs/list`` then fetches ``/api/runs/shared/{hash}``
for every run not already on disk, writing the JSON to
``<data>/files-spire-codex/<character>/<YYYY-MM>/<hash>.run``.

The listing endpoint is the bottleneck (~10s/page, capped at 50 rows/page), so
pages are fetched concurrently rather than one at a time — sequential paging of
the full corpus took *hours*. Detail fetches are fast (~0.3s) and also run
concurrently. Runs already on disk, and run_hashes recorded as permanently