chatmock/routes_custom.py

from __future__ import annotations

import json
from flask import Blueprint, jsonify, request, current_app
from .routes_openai import chat_completions

custom_bp = Blueprint("custom", __name__)

@custom_bp.route("/api", methods=["POST"])
@custom_bp.route("/model/api", methods=["POST"])
@custom_bp.route("/<model_name>/api", methods=["POST"])
def custom_api(model_name: str | None = None):
    raw = request.get_data(as_text=True) or ""
    try:
        payload = json.loads(raw)
    except Exception:
        return jsonify({"status": "error", "message": "Invalid JSON"}), 400
    
    prompt = payload.get("prompt")
    if not prompt:
        return jsonify({"status": "error", "message": "Missing prompt"}), 400

    from .upstream import normalize_model_name, start_upstream_request
    from .responses_api import instructions_for_model
    from .reasoning import build_reasoning_param, allowed_efforts_for_model
    
    # Use model_name from URL if provided, otherwise default to config or None
    model = normalize_model_name(model_name, current_app.config.get("DEBUG_MODEL"))
    
    input_items = [
        {"type": "message", "role": "user", "content": [{"type": "input_text", "text": prompt}]}
    ]
    
    reasoning_param = build_reasoning_param(
        current_app.config.get("REASONING_EFFORT", "medium"),
        current_app.config.get("REASONING_SUMMARY", "auto"),
        None,
        allowed_efforts=allowed_efforts_for_model(model)
    )
    
    instructions = instructions_for_model(current_app.config, model)
    
    upstream, error_resp = start_upstream_request(
        model,
        input_items,
        instructions=instructions,
        reasoning_param=reasoning_param
    )
    
    if error_resp:
        return jsonify({"status": "error", "message": "Upstream error"}), 502
        
    if upstream.status_code >= 400:
        return jsonify({"status": "error", "message": f"Upstream returned {upstream.status_code}"}), upstream.status_code

    full_text = ""
    try:
        for raw_line in upstream.iter_lines(decode_unicode=False):
            if not raw_line: continue
            line = raw_line.decode("utf-8", errors="ignore")
            if not line.startswith("data: "): continue
            data = line[len("data: "):].strip()
            if not data or data == "[DONE]": break
            try:
                evt = json.loads(data)
                if evt.get("type") == "response.output_text.delta":
                    full_text += evt.get("delta") or ""
                elif evt.get("type") == "response.completed":
                    break
            except Exception:
                continue
    finally:
        upstream.close()

    # Clean the text: remove markdown, special unicode, and newlines
    clean_text = full_text.strip()
    
    # 1. Handle common special unicode characters
    replacements = {
        "\u2019": "'",
        "\u2018": "'",
        "\u201d": '"',
        "\u201c": '"',
        "\u2014": "-",
        "\u2013": "-",
        "\u2022": "*",
        "\u2026": "..."
    }
    for old, new in replacements.items():
        clean_text = clean_text.replace(old, new)
        
    # 2. Strip basic markdown (bold, italics)
    import re
    clean_text = re.sub(r'\*\*(.*?)\*\*', r'\1', clean_text) # bold
    clean_text = re.sub(r'\*(.*?)\*', r'\1', clean_text)     # italics
    clean_text = re.sub(r'__(.*?)__', r'\1', clean_text)     # bold
    clean_text = re.sub(r'_(.*?)_', r'\1', clean_text)       # italics
    
    # 3. Handle newlines: replace with spaces
    clean_text = clean_text.replace("\n\n", " ").replace("\n", " ")
    
    # 4. Remove emojis and other special non-ascii characters for clean plain text
    # This removes the \ud83d\ude04 type sequences
    clean_text = clean_text.encode('ascii', 'ignore').decode('ascii')
    
    # 5. Final trim
    clean_text = re.sub(r'\s+', ' ', clean_text).strip()

    return jsonify({
        "status": "success",
        "text": clean_text
    })