feat: add pg_documentdb and pg_cron extensions to sql image

Multi-stage Debian build with pgvector, pg_cron, and pg_documentdb
(core + api). Enables FerretDB/docdb MongoDB compatibility and
scheduled job support. PostGIS included as documentdb runtime dep.

Author: hanzo-dev

Dockerfile

@@ -1,38 +1,128 @@
 ARG PG_MAJOR=18
 
-# Stage 1: Build pgvector extension from Hanzo fork
-FROM postgres:${PG_MAJOR}-alpine AS builder
+# Stage 1: Build all extensions on Debian (documentdb requires glibc toolchain)
+FROM postgres:${PG_MAJOR}-bookworm AS builder
 ARG PG_MAJOR
 
-RUN apk add --no-cache --virtual .build-deps \
-        build-base \
-        clang19 \
-        llvm19 \
-        postgresql${PG_MAJOR}-dev && \
-    cd /tmp && \
-    wget -qO- https://github.com/hanzoai/sql-vector/archive/refs/heads/master.tar.gz | tar xz && \
+ENV DEBIAN_FRONTEND=noninteractive
+
+# Install build dependencies for pgvector, pg_cron, and pg_documentdb
+RUN apt-get update && apt-get install -y --no-install-recommends \
+        build-essential \
+        cmake \
+        git \
+        curl \
+        ca-certificates \
+        pkg-config \
+        postgresql-server-dev-${PG_MAJOR} \
+        libpq-dev \
+        libicu-dev \
+        libkrb5-dev \
+        # PostGIS (runtime dep for documentdb CREATE EXTENSION)
+        postgresql-${PG_MAJOR}-postgis-3 \
+    && rm -rf /var/lib/apt/lists/*
+
+# --- pgvector ---
+RUN cd /tmp && \
+    curl -sL https://github.com/hanzoai/sql-vector/archive/refs/heads/master.tar.gz | tar xz && \
     cd sql-vector-master && \
     make clean && \
-    make OPTFLAGS="" PG_CONFIG=/usr/local/bin/pg_config && \
-    make install PG_CONFIG=/usr/local/bin/pg_config && \
-    cd / && rm -rf /tmp/sql-vector-master && \
-    apk del .build-deps
+    make OPTFLAGS="" PG_CONFIG=/usr/bin/pg_config && \
+    make install PG_CONFIG=/usr/bin/pg_config && \
+    cd / && rm -rf /tmp/sql-vector-master
+
+# --- pg_cron ---
+ARG PG_CRON_VERSION=v1.6.7
+RUN cd /tmp && \
+    curl -sL https://github.com/citusdata/pg_cron/archive/refs/tags/${PG_CRON_VERSION}.tar.gz | tar xz && \
+    cd pg_cron-* && \
+    make PG_CONFIG=/usr/bin/pg_config && \
+    make install PG_CONFIG=/usr/bin/pg_config && \
+    cd / && rm -rf /tmp/pg_cron-*
+
+# --- pg_documentdb dependencies: libbson, Intel Decimal Math Lib, pcre2 ---
+
+ARG LIBBSON_VERSION=1.28.0
+RUN cd /tmp && \
+    curl -sL https://github.com/mongodb/mongo-c-driver/releases/download/${LIBBSON_VERSION}/mongo-c-driver-${LIBBSON_VERSION}.tar.gz | tar xz && \
+    cd mongo-c-driver-${LIBBSON_VERSION}/build && \
+    cmake -DENABLE_MONGOC=ON -DMONGOC_ENABLE_ICU=OFF -DENABLE_ICU=OFF \
+          -DCMAKE_C_FLAGS="-fPIC -g" -DCMAKE_BUILD_TYPE=Release \
+          -DCMAKE_INSTALL_PREFIX=/usr .. && \
+    make -j$(nproc) install && \
+    cd / && rm -rf /tmp/mongo-c-driver-*
+
+ARG INTEL_MATH_LIB_VERSION=applied/2.0u3-1
+RUN cd /tmp && \
+    mkdir -p intelmathlib/lib/intelmathlib && \
+    cd intelmathlib/lib/intelmathlib && \
+    git init && \
+    git remote add origin https://git.launchpad.net/ubuntu/+source/intelrdfpmath && \
+    git fetch --depth 1 origin "${INTEL_MATH_LIB_VERSION}" && \
+    git checkout FETCH_HEAD && \
+    cd LIBRARY && \
+    make -j$(nproc) _CFLAGS_OPT=-fPIC CC=gcc CALL_BY_REF=0 GLOBAL_RND=0 GLOBAL_FLAGS=0 UNCHANGED_BINARY_FLAGS=0 && \
+    cd /tmp/intelmathlib && \
+    # Create pkg-config file
+    printf 'prefix=/usr/lib/intelmathlib\nlibdir=${prefix}/LIBRARY\nincludedir=${prefix}/LIBRARY/src\n\nName: intelmathlib\nDescription: Intel Decimal Floating point math library\nVersion: 2.0 Update 2\nCflags: -I${includedir}\nLibs: -L${libdir} -lbid\n' > intelmathlib.pc && \
+    cp -R lib/intelmathlib /usr/lib/ && \
+    cp intelmathlib.pc /usr/lib/pkgconfig/ && \
+    cd / && rm -rf /tmp/intelmathlib
+
+ARG PCRE2_VERSION=10.40
+RUN cd /tmp && \
+    curl -sL https://github.com/PCRE2Project/pcre2/releases/download/pcre2-${PCRE2_VERSION}/pcre2-${PCRE2_VERSION}.tar.gz | tar xz && \
+    cd pcre2-${PCRE2_VERSION} && \
+    ./configure --prefix=/usr --disable-shared --enable-static --enable-jit && \
+    make -j$(nproc) AM_CFLAGS=-fPIC install && \
+    cd / && rm -rf /tmp/pcre2-*
 
-# Stage 2: Runtime image
-FROM postgres:${PG_MAJOR}-alpine
+# --- pg_documentdb (core + api) ---
+ARG DOCUMENTDB_VERSION=main
+RUN cd /tmp && \
+    curl -sL https://github.com/microsoft/documentdb/archive/refs/heads/${DOCUMENTDB_VERSION}.tar.gz | tar xz && \
+    cd documentdb-* && \
+    make -C pg_documentdb_core PG_CONFIG=/usr/bin/pg_config install && \
+    make -C pg_documentdb PG_CONFIG=/usr/bin/pg_config install && \
+    cd / && rm -rf /tmp/documentdb-*
+
+# Stage 2: Runtime image (Debian for glibc compat with documentdb)
+FROM postgres:${PG_MAJOR}-bookworm
 
 LABEL maintainer="dev@hanzo.ai"
 LABEL org.opencontainers.image.source="https://github.com/hanzoai/sql"
-LABEL org.opencontainers.image.description="Hanzo SQL - PostgreSQL with pgvector for AI workloads"
+LABEL org.opencontainers.image.description="Hanzo SQL - PostgreSQL with pgvector, pg_cron, and pg_documentdb"
+
+# Runtime deps: libpq (pg_cron), PostGIS (documentdb geospatial), ICU
+ARG PG_MAJOR=18
+RUN apt-get update && apt-get install -y --no-install-recommends \
+        libpq5 \
+        postgresql-${PG_MAJOR}-postgis-3 \
+    && rm -rf /var/lib/apt/lists/*
+
+# Copy Intel Decimal Math Library (runtime dep for documentdb)
+COPY --from=builder /usr/lib/intelmathlib /usr/lib/intelmathlib
+
+# Copy pgvector extension
+COPY --from=builder /usr/lib/postgresql/${PG_MAJOR}/lib/vector.so /usr/lib/postgresql/${PG_MAJOR}/lib/
+COPY --from=builder /usr/share/postgresql/${PG_MAJOR}/extension/vector* /usr/share/postgresql/${PG_MAJOR}/extension/
+
+# Copy pg_cron extension
+COPY --from=builder /usr/lib/postgresql/${PG_MAJOR}/lib/pg_cron.so /usr/lib/postgresql/${PG_MAJOR}/lib/
+COPY --from=builder /usr/share/postgresql/${PG_MAJOR}/extension/pg_cron* /usr/share/postgresql/${PG_MAJOR}/extension/
+
+# Copy pg_documentdb_core extension
+COPY --from=builder /usr/lib/postgresql/${PG_MAJOR}/lib/pg_documentdb_core.so /usr/lib/postgresql/${PG_MAJOR}/lib/
+COPY --from=builder /usr/share/postgresql/${PG_MAJOR}/extension/documentdb_core* /usr/share/postgresql/${PG_MAJOR}/extension/
 
-# Copy pgvector extension from builder (shared object + SQL files)
-COPY --from=builder /usr/local/lib/postgresql/vector.so /usr/local/lib/postgresql/
-COPY --from=builder /usr/local/share/postgresql/extension/vector* /usr/local/share/postgresql/extension/
+# Copy pg_documentdb extension
+COPY --from=builder /usr/lib/postgresql/${PG_MAJOR}/lib/pg_documentdb.so /usr/lib/postgresql/${PG_MAJOR}/lib/
+COPY --from=builder /usr/share/postgresql/${PG_MAJOR}/extension/documentdb* /usr/share/postgresql/${PG_MAJOR}/extension/
 
 # Custom postgresql.conf tuned for AI workloads
 COPY conf/postgresql.conf /etc/postgresql/postgresql.conf
 
-# Init scripts: enable pgvector on startup
+# Init scripts: enable extensions on startup
 COPY docker-entrypoint-initdb.d/ /docker-entrypoint-initdb.d/
 
 # Health check

conf/postgresql.conf

@@ -45,7 +45,10 @@ lc_numeric = 'en_US.utf8'
 lc_time = 'en_US.utf8'
 
 # Extensions
-shared_preload_libraries = 'vector'
+shared_preload_libraries = 'vector,pg_cron,pg_documentdb_core,pg_documentdb'
+
+# pg_cron - run jobs in any database (not just postgres)
+cron.database_name = 'postgres'
 
 # pgvector specific - HNSW index build parallelism
 max_parallel_maintenance_workers = 4

docker-entrypoint-initdb.d/01-extensions.sql

@@ -1,2 +1,7 @@
--- Hanzo SQL: Enable pgvector extension
+-- Hanzo SQL: Enable extensions
 CREATE EXTENSION IF NOT EXISTS vector;
+CREATE EXTENSION IF NOT EXISTS pg_cron;
+CREATE EXTENSION IF NOT EXISTS postgis;
+CREATE EXTENSION IF NOT EXISTS tsm_system_rows;
+CREATE EXTENSION IF NOT EXISTS documentdb_core;
+CREATE EXTENSION IF NOT EXISTS documentdb;