Dockerfile.data

# Base working stage
FROM node:22-slim AS base
ENV PNPM_HOME="/pnpm"
ENV PATH="$PNPM_HOME:$PATH"
RUN corepack enable
WORKDIR /app

###################
## Data Stage — download ENS Subgraph rainbow table archive & checksum
###################

FROM base AS db-data
RUN apt-get update && apt-get install -y wget
COPY apps/ensrainbow/download-rainbow-tables.sh .

# Allow parametrization of which data version to download
# Possible values: "v1" (default) or "v2" or "test"
# This ARG is used to determine which data fetching command to run during the build process.
# It can be set using the --build-arg flag when building the Docker image.
ARG DATA_VERSION="v1"
RUN DATA_VERSION=${DATA_VERSION} ./download-rainbow-tables.sh

###################
## Ingest Stage — ingest to produce data dir artifact
###################

FROM base AS ingest
COPY . .
RUN --mount=type=cache,id=pnpm,target=/pnpm/store pnpm install --frozen-lockfile --ignore-scripts

WORKDIR /app/apps/ensrainbow
# copy ens_names.sql.gz from db-data
COPY --from=db-data /app/ens_names.sql.gz /app/THE_GRAPH_LICENSE.txt .

# produce the data dir artifact
RUN pnpm run ingest

# cleanup input file to reduce layer size
RUN rm ens_names.sql.gz

###################
## Output Image — just the data dir
###################

FROM base AS ensrainbow-data
ARG DATA_VERSION="v1"
# Set a label to identify the data version
LABEL io.ensnode.data-version=${DATA_VERSION}

COPY --from=ingest /app/apps/ensrainbow/data ./data