pdf/gateway/app/routers/render.py

import hashlib
from typing import Annotated

from fastapi import APIRouter, HTTPException, Path, Query, Request, Response, status

from app.schemas.font import FontInfoResponse
from app.services import engine
from app.services.font import font_info_to_response
from app.services.render_cache import RENDERER_VERSION, RenderMode, TileCacheKey, tile_cache
from app.services.store import document_store

router = APIRouter(prefix="/documents/{document_id}/pages", tags=["render"])
compat_router = APIRouter(tags=["render"])

def _etag(data: bytes) -> str:
    return hashlib.sha256(data).hexdigest()[:16]


@router.get("/{page_index}/render")
def render_page(
    request: Request,
    document_id: str, 
    page_index: Annotated[int, Path(ge=0)], 
    dpi: int = 96,
    zoom: float = 1.0,
    rotation: int = 0,
    render_mode: RenderMode = RenderMode.NORMAL
) -> Response:
    print(f"[RENDER] document_id={document_id} page_index={page_index} zoom={zoom} dpi={dpi}")
    if not engine.is_available():
        raise HTTPException(
            status_code=status.HTTP_501_NOT_IMPLEMENTED,
            detail="Engine bridge (bindings/python) not yet available.",
        )

    doc_info = document_store.get_document(document_id)
    if not doc_info:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Document not found")

    try:
        doc = doc_info["doc_instance"]
        page = doc.get_page(page_index)
        
        key = TileCacheKey(
            doc_hash=doc_info.get("doc_hash", ""),
            renderer_version=RENDERER_VERSION,
            page=page_index,
            dpi=dpi,
            zoom=zoom,
            rotation=rotation,
            render_mode=render_mode,
            tile_x=0.0,
            tile_y=0.0,
            tile_w=round(page.width, 2),
            tile_h=round(page.height, 2)
        )
        
        cached = tile_cache.get(key)
        if cached:
            etag_val = f'"{_etag(cached)}"'
            if request.headers.get("if-none-match") == etag_val:
                return Response(status_code=304)
            return Response(
                content=cached, 
                media_type="image/png",
                headers={"ETag": etag_val, "X-Cache": "HIT", "Cache-Control": "private, max-age=300"}
            )
            
        import time
        start = time.perf_counter_ns()
        img = page.render(dpi)
        elapsed = time.perf_counter_ns() - start
        
        tile_cache.put(key, img.data)
        tile_cache.record_render_time(elapsed)
        
        etag_val = f'"{_etag(img.data)}"'
        return Response(
            content=img.data, 
            media_type="image/png",
            headers={"ETag": etag_val, "X-Cache": "MISS", "Cache-Control": "private, max-age=300"}
        )
    except IndexError:
        raise HTTPException(
            status_code=status.HTTP_404_NOT_FOUND, detail="Page index out of bounds"
        )
    except Exception as e:
        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))


@router.get("/{page_index}/render-tile")
def render_page_tile(
    request: Request,
    document_id: str,
    page_index: Annotated[int, Path(ge=0)],
    x: float,
    y: float,
    width: float,
    height: float,
    dpi: int = 96,
    zoom: float = 1.0,
    rotation: int = 0,
    render_mode: RenderMode = RenderMode.NORMAL
) -> Response:
    if not engine.is_available():
        raise HTTPException(
            status_code=status.HTTP_501_NOT_IMPLEMENTED,
            detail="Engine bridge (bindings/python) not yet available.",
        )

    doc_info = document_store.get_document(document_id)
    if not doc_info:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Document not found")

    try:
        doc = doc_info["doc_instance"]
        page = doc.get_page(page_index)
        
        key = TileCacheKey(
            doc_hash=doc_info.get("doc_hash", ""),
            renderer_version=RENDERER_VERSION,
            page=page_index,
            dpi=dpi,
            zoom=zoom,
            rotation=rotation,
            render_mode=render_mode,
            tile_x=round(x, 2),
            tile_y=round(y, 2),
            tile_w=round(width, 2),
            tile_h=round(height, 2)
        )
        
        cached = tile_cache.get(key)
        if cached:
            etag_val = f'"{_etag(cached)}"'
            if request.headers.get("if-none-match") == etag_val:
                return Response(status_code=304)
            return Response(
                content=cached, 
                media_type="image/png",
                headers={"ETag": etag_val, "X-Cache": "HIT", "Cache-Control": "private, max-age=300"}
            )

        import time
        start = time.perf_counter_ns()
        img_width, img_height, img_data = page.render_tile(dpi, x, y, width, height)
        import io
        from PIL import Image
        img = Image.frombytes("RGBA", (img_width, img_height), img_data)
        out_buf = io.BytesIO()
        img.save(out_buf, format="PNG")
        png_bytes = out_buf.getvalue()
        elapsed = time.perf_counter_ns() - start
        
        tile_cache.put(key, png_bytes)
        tile_cache.record_render_time(elapsed)
        
        etag_val = f'"{_etag(png_bytes)}"'
        return Response(
            content=png_bytes, 
            media_type="image/png",
            headers={"ETag": etag_val, "X-Cache": "MISS", "Cache-Control": "private, max-age=300"}
        )
    except IndexError:
        raise HTTPException(
            status_code=status.HTTP_404_NOT_FOUND, detail="Page index out of bounds"
        )
    except Exception as e:
        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))


@router.get("/{page_index}/text")
def extract_page_text(document_id: str, page_index: Annotated[int, Path(ge=0)]):
    if not engine.is_available():
        raise HTTPException(
            status_code=status.HTTP_501_NOT_IMPLEMENTED,
            detail="Engine bridge (bindings/python) not yet available.",
        )

    doc_info = document_store.get_document(document_id)
    if not doc_info:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Document not found")

    try:
        doc = doc_info["doc_instance"]
        page = doc.get_page(page_index)
        text = page.extract_text()
        annots = page.extract_annotations_text()
        if annots:
            text += "\n" + "\n".join(annots)
        page_height = page.height
        glyphs = page.extract_text_with_bounds()
        for g in glyphs:
            g["y"] = page_height - (g["y"] + g["h"])
        return {"text": text, "glyphs": glyphs}
    except IndexError:
        raise HTTPException(
            status_code=status.HTTP_404_NOT_FOUND, detail="Page index out of bounds"
        )
    except Exception as e:
        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))


@compat_router.get("/render/{document_id}")
def render_page_compat(
    request: Request, document_id: str, page: int = 0, zoom: float = 1.0, rotation: int = 0, dpi: int | None = None
) -> Response:
    if dpi is None:
        dpi = int(96 * zoom)
    print(f"[RENDER_COMPAT] document_id={document_id} page={page} zoom={zoom} dpi={dpi}")
    return render_page(request, document_id, page, dpi, zoom, rotation, RenderMode.NORMAL)


@router.get("/{page_index}")
def get_page_info(document_id: str, page_index: Annotated[int, Path(ge=0)]):
    if not engine.is_available():
        raise HTTPException(
            status_code=status.HTTP_501_NOT_IMPLEMENTED,
            detail="Engine bridge (bindings/python) not yet available.",
        )

    doc_info = document_store.get_document(document_id)
    if not doc_info:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Document not found")

    try:
        doc = doc_info["doc_instance"]
        page = doc.get_page(page_index)
        return {"width": page.width, "height": page.height}
    except IndexError:
        raise HTTPException(
            status_code=status.HTTP_404_NOT_FOUND, detail="Page index out of bounds"
        )
    except Exception as e:
        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))


@router.get("/{page_index}/transform/page-to-device")
def transform_page_to_device(
    document_id: str,
    page_index: int,
    x: float,
    y: float,
    device_width: int,
    device_height: int,
    rotate: int = 0,
):
    if not engine.is_available():
        raise HTTPException(
            status_code=status.HTTP_501_NOT_IMPLEMENTED, detail="Engine unavailable"
        )

    doc_info = document_store.get_document(document_id)
    if not doc_info:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Document not found")

    try:
        pdfengine = engine.require()
        doc = doc_info["doc_instance"]
        page = doc.get_page(page_index)

        pt = pdfengine.Point2D(x=x, y=y)
        res = page.page_to_device(pt, device_width, device_height, rotate)
        return {"x": res.x, "y": res.y}
    except Exception as e:
        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))


@router.get("/{page_index}/transform/device-to-page")
def transform_device_to_page(
    document_id: str,
    page_index: int,
    x: int,
    y: int,
    device_width: int,
    device_height: int,
    rotate: int = 0,
):
    if not engine.is_available():
        raise HTTPException(
            status_code=status.HTTP_501_NOT_IMPLEMENTED, detail="Engine unavailable"
        )

    doc_info = document_store.get_document(document_id)
    if not doc_info:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Document not found")

    try:
        pdfengine = engine.require()
        doc = doc_info["doc_instance"]
        page = doc.get_page(page_index)

        pt = pdfengine.DevicePoint(x=x, y=y)
        res = page.device_to_page(pt, device_width, device_height, rotate)
        return {"x": res.x, "y": res.y}
    except Exception as e:
        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))


@router.get("/{page_index}/fonts", response_model=list[FontInfoResponse])
def get_page_fonts(document_id: str, page_index: int) -> list[FontInfoResponse]:
    if not engine.is_available():
        raise HTTPException(
            status_code=status.HTTP_501_NOT_IMPLEMENTED,
            detail="Engine bridge (bindings/python) not yet available.",
        )

    doc_info = document_store.get_document(document_id)
    if not doc_info:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Document not found")

    try:
        doc = doc_info["doc_instance"]
        page = doc.get_page(page_index)
        fonts = page.get_fonts()
        return [font_info_to_response(f) for f in fonts]
    except IndexError:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Page index out of range")
    except Exception as e:
        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))


@router.get("/{page_index}/fonts/glyph-width")
def get_page_glyph_width(
    document_id: str,
    page_index: Annotated[int, Path(ge=0)],
    font_name: str,
    charcode: int,
    font_size: Annotated[float, Query(gt=0)] = 12.0,
):
    if not engine.is_available():
        raise HTTPException(
            status_code=status.HTTP_501_NOT_IMPLEMENTED,
            detail="Engine bridge (bindings/python) not yet available.",
        )

    doc_info = document_store.get_document(document_id)
    if not doc_info:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Document not found")

    try:
        doc = doc_info["doc_instance"]
        page = doc.get_page(page_index)
        width = page.get_glyph_width(font_name, charcode, font_size)
        return {"width": width}
    except IndexError:
        raise HTTPException(status_code=status.HTTP_404_NOT_FOUND, detail="Page index out of range")
    except Exception as e:
        raise HTTPException(status_code=status.HTTP_400_BAD_REQUEST, detail=str(e))