#!/usr/bin/env bash
#
# AOCore streaming chat-completion via curl (SSE).
#
# Usage:
#   export AOSENTRY_KEY="sk-..."
#   bash streaming.sh
#
# CRITICAL: curl's -N flag disables output buffering. Without -N, curl waits
# until the full response is received before printing anything, which defeats
# the entire point of SSE. Always pass -N for streaming endpoints.
#
# AOCore stamps the six X-AOCore-* quota headers BEFORE the first SSE
# chunk is emitted (proven by TestStreamWriter_StampsHeadersBeforeWriteHeader
# in go/internal/llm/streaming_quota_headers_test.go), so the quota state is
# visible in the response headers — see ../../rate-limits.md.
set -euo pipefail

: "${AOSENTRY_KEY:?Set AOSENTRY_KEY to your minted API key (sk-...).}"

BASE_URL="${AOSENTRY_BASE_URL:-https://gateway.core.aocyber.ai}"

curl -N "${BASE_URL}/v1/chat/completions" \
  -H "Authorization: Bearer ${AOSENTRY_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [
      {"role": "user", "content": "Count from 1 to 5, one number per line."}
    ],
    "stream": true
  }'