File size: 4,872 Bytes
2fd8593
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8f2f756
 
 
 
 
2fd8593
 
 
 
8f2f756
2fd8593
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8f2f756
2fd8593
 
8f2f756
2fd8593
 
 
 
 
 
 
 
 
 
 
 
 
8f2f756
2fd8593
 
 
 
 
 
 
 
8f2f756
2fd8593
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
import os, sys, shutil, tempfile, zipfile, asyncio, subprocess
from pathlib import Path
from fastapi import FastAPI, UploadFile, File, Form, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import FileResponse

REPO_ROOT = Path(__file__).parent.resolve()
CODES_DIR = REPO_ROOT / "codes"

app = FastAPI(title="Blog2Code API", version="1.0.0")
ALLOWED_ORIGINS = os.getenv("ALLOWED_ORIGINS", "*").split(",")
app.add_middleware(
    CORSMiddleware,
    allow_origins=ALLOWED_ORIGINS,
    allow_methods=["*"],
    allow_headers=["*"],
)

def _run(script: str, args: list, extra_env: dict) -> None:
    cmd = [sys.executable, str(CODES_DIR / script)] + args
    result = subprocess.run(
        cmd,
        cwd=str(REPO_ROOT),
        env={**os.environ, **extra_env},
        capture_output=True,
        text=True,
    )
    if result.returncode != 0:
        raise RuntimeError(
            f"{script} failed (exit {result.returncode}):\n"
            f"STDOUT: {result.stdout[-2000:]}\n"
            f"STDERR: {result.stderr[-2000:]}"
        )

@app.get("/health")
def health():
    return {"status": "ok"}

@app.post("/generate")
async def generate(
    url:  str        = Form(None),
    file: UploadFile = File(None),
):
    if not url and not file:
        raise HTTPException(400, "Provide either 'url' or 'file'.")

    tmp        = Path(tempfile.mkdtemp())
    data_dir   = tmp / "data"
    output_dir = tmp / "output"
    data_dir.mkdir(parents=True)
    output_dir.mkdir(parents=True)

    try:
        if file:
            suffix     = Path(file.filename).suffix or ".md"
            input_path = tmp / f"blog{suffix}"
            input_path.write_bytes(await file.read())
            source_args = ["--input_path", str(input_path)]
        else:
            source_args = ["--url", url.strip()]

        provider = os.getenv("PROVIDER", "gemma")
        # Default model for NVIDIA/Llama β€” overridable via MODEL env var
        default_model = "meta/llama-3.3-70b-instruct"
        model = os.getenv("MODEL", default_model)
        extra_env = {"MODEL": model}

        blog_json = data_dir / "blog_data.json"

        def run_pipeline():
            # Stage 0 – parse blog (no LLM, no --model needed)
            _run("0_blog_process.py",
                 source_args + ["--output_json_path", str(blog_json)],
                 extra_env)

            if not blog_json.exists():
                candidates = list(data_dir.glob("*.json"))
                if not candidates:
                    raise RuntimeError("Stage 0: no JSON output found.")
                blog_json_path = candidates[0]
            else:
                blog_json_path = blog_json

            # Stage 1 – planning
            _run("1_planning.py", [
                "--blog_json_path", str(blog_json_path),
                "--output_dir",     str(data_dir),
                "--provider",       provider,
                "--content_type",   "blog",
                "--model",          model,
            ], extra_env)

            # Stage 1.1 – extract config (no LLM, no --model needed)
            _run("1_1_extract_config.py", [
                "--output_dir", str(data_dir),
            ], extra_env)

            config_yaml = data_dir / "planning_config.yaml"
            if not config_yaml.exists():
                raise RuntimeError("Stage 1.1: planning_config.yaml not found.")

            # Stage 2 – analysis
            _run("2_analyzing.py", [
                "--pdf_json_path", str(blog_json_path),
                "--output_dir",    str(data_dir),
                "--provider",      provider,
                "--model",         model,
            ], extra_env)

            # Stage 3 – code generation
            _run("3_coding.py", [
                "--pdf_json_path",   str(blog_json_path),
                "--output_dir",      str(data_dir),
                "--output_repo_dir", str(output_dir),
                "--provider",        provider,
                "--model",           model,
            ], extra_env)

        await asyncio.get_event_loop().run_in_executor(None, run_pipeline)

        zip_path = tmp / "repo.zip"
        files = [f for f in output_dir.rglob("*") if f.is_file()]
        if not files:
            raise HTTPException(500, "Pipeline produced no output files.")

        with zipfile.ZipFile(zip_path, "w", zipfile.ZIP_DEFLATED) as zf:
            for f in files:
                zf.write(f, f.relative_to(output_dir))

        return FileResponse(
            path=str(zip_path),
            media_type="application/zip",
            filename="generated-repo.zip",
        )

    except HTTPException:
        shutil.rmtree(tmp, ignore_errors=True)
        raise
    except Exception as exc:
        shutil.rmtree(tmp, ignore_errors=True)
        raise HTTPException(500, str(exc)) from exc