docling v1 API changes break OCR requests

[darkorb]

Hi there,

I've not familiar enough with the code here to fix this up myself, but noticed that the API being used in here for docling has been migrated from /v1alpha/ with some changes to how it's used.

https://github.com/docling-project/docling-serve/blob/main/docs/v1_migration.md#endpoint-names

Currently if you pull a recent docling image this will result in OCR via it breaking. A workaround for now (apparently) being to use the version tagged v0.16.1 in your image pulls.

[inboxcda]

also dealing with this issue

[inboxcda]

I fixed this by downloading the repo and changing the api paths in the code to us v1 instead of v1alpha. While it works, I am having some timeout issues with docling and I am not sure if they're related to me using a non-official and hacked version since I changed the api paths to docling.... looking forward to getting an enhancement on the docling integration so I can rule out this other issue.

You can see in my logs at the bottom of the paperless-gpt section that an api call to docling was made but docling through a warnings or two for the docling server logs section and it seems to relate to that api call. Then you see nothing else happens but health check pings which I setup. Eventually docling throws a timeout error of sorts and I am not sure how to resolve it. I m waiting for that screenshot and will post it shortly.

[T-Eberle]

@inboxcda I have done the same, I will add a PR this evening including also some other small docling changes (adding DOCLING_OCR_PIPELINE and DOCLING_OCR_ENGINE, as I prefer using easyocr instead of vlms...)

Referring to your timeouts: If docling throws a timeout (I can't see them in your screenshots), then you will probably have to set the environment variable DOCLING_SERVE_MAX_SYNC_WAIT (default to 120)

[T-Eberle]

Fixed in #614

[inboxcda]

I am still having this issue though with latest pull. I assume I have to wait for your pull request to be accepted for it to be incorporated into the paperless-gpt image?

@T-Eberle Thank you for working on this!!

[inboxcda]

I checked out your pr 614 and built a Dockerfile with it but I am struggling to get it working. I will keep trying later tonight or this week. Thanks again!

ARG VERSION=pr-614-dev
ARG COMMIT=pr-614
ARG BUILD_DATE=unknown

# Stage 0: Fetch PR #614 source code
FROM docker.io/alpine:3.22.1 AS source
RUN apk add --no-cache git
WORKDIR /src
# RUN git clone https://github.com/icereed/paperless-gpt.git .
# RUN git fetch origin pull/614/head:pr-614
# RUN git checkout pr-614
RUN git clone https://github.com/T-Eberle/paperless-gpt.git .

# Stage 1: Build Vite frontend
FROM docker.io/node:22-alpine AS frontend
# Set the working directory inside the container
WORKDIR /app
# Install necessary packages
RUN apk add --no-cache git
# Copy package.json and package-lock.json from PR source
COPY --from=source /src/web-app/package.json /src/web-app/package-lock.json ./
# Install dependencies
RUN npm install
# Copy the frontend code from PR source
COPY --from=source /src/web-app /app/
# Build the frontend
RUN npm run build

# Stage 2: Build the Go binary
FROM docker.io/golang:1.25.1-alpine3.21 AS builder
# Set the working directory inside the container
WORKDIR /app
# Package versions for Renovate
# renovate: datasource=repology depName=alpine_3_21/gcc versioning=loose
ENV GCC_VERSION="14.2.0-r4"
# renovate: datasource=repology depName=alpine_3_21/musl-dev versioning=loose
ENV MUSL_DEV_VERSION="1.2.5-r9"
# renovate: datasource=repology depName=alpine_3_21/mupdf versioning=loose
ENV MUPDF_VERSION="1.24.10-r0"
# renovate: datasource=repology depName=alpine_3_21/mupdf-dev versioning=loose
ENV MUPDF_DEV_VERSION="1.24.10-r0"
# renovate: datasource=repology depName=alpine_3_21/sed versioning=loose
ENV SED_VERSION="4.9-r2"
# Install necessary packages with pinned versions
RUN apk add --no-cache \
    "gcc=${GCC_VERSION}" \
    "musl-dev=${MUSL_DEV_VERSION}" \
    "mupdf=${MUPDF_VERSION}" \
    "mupdf-dev=${MUPDF_DEV_VERSION}" \
    "sed=${SED_VERSION}"
# Copy go.mod and go.sum files from PR source
COPY --from=source /src/go.mod /src/go.sum ./
# Download dependencies
RUN go mod download
# Pre-compile go-sqlite3 to avoid doing this every time
RUN CGO_ENABLED=1 go build -tags musl -o /dev/null github.com/mattn/go-sqlite3
# Copy the frontend build
COPY --from=frontend /app/dist /app/web-app/dist
# Copy the Go source files from PR source
COPY --from=source /src/*.go ./
COPY --from=source /src/ocr ./ocr
# Import ARGs from top level
ARG VERSION
ARG COMMIT
ARG BUILD_DATE
# Update version information
RUN sed -i \
    -e "s/devVersion/${VERSION}/" \
    -e "s/devBuildDate/${BUILD_DATE}/" \
    -e "s/devCommit/${COMMIT}/" \
    version.go
# Build the binary using caching for both go modules and build cache
RUN CGO_ENABLED=1 GOMAXPROCS=$(nproc) go build -tags musl -o paperless-gpt .

# Stage 3: Create a lightweight image with just the binary
FROM docker.io/alpine:3.22.1
ENV GIN_MODE=release
# Install necessary runtime dependencies
RUN apk add --no-cache \
    ca-certificates
# Set the working directory inside the container
WORKDIR /app/
# Copy the Go binary from the builder stage
COPY --from=builder /app/paperless-gpt .
# Copy the prompt templates from PR source
COPY --from=source /src/default_prompts/ /app/default_prompts/
# Expose the port the app runs on
EXPOSE 8080
# Command to run the binary
CMD ["/app/paperless-gpt"]

  paperless-gpt:
    container_name: paperless-gpt

    # Use one of these image sources:
    # image: icereed/paperless-gpt:latest # Docker Hub
    # image: ghcr.io/icereed/paperless-gpt:latest  # GitHub Container Registry
    # Watchtower configuration
    # com.centurylinklabs.watchtower.enable: "false"
    # labels:
    #   com.centurylinklabs.watchtower.enable: "false"


    # Comment/uncomment these lines to switch between versions
    # image: paperless-gpt:commit-614          # Your custom build
    # image: icereed/paperless-gpt:latest    # Official image

    build:
      context: .
      dockerfile: Dockerfile

[T-Eberle]

Hi @inboxcda ,

you need to download the models: I recommend downloading everything when using docling to be on the absolute safe side ;)

I recommend the following steps:

• Define a volume for your docker container running docling-serve: docker run --rm --name loader -v #YOUR_HOST_PATH:/models ghcr.io/docling-project/docling-serve-cpu:main docling-tools models download --output-dir=/models --all
• Run the docker container docling-serve and mounting the created volume with the models inside the docling-serve container.
• Add the environment variable DOCLING_SERVE_ARTIFACTS_PATH with the volume container path inside the docling-serve container.

I hope that helps.