Add Python to docker containers

crates/arroyo-connectors/Cargo.toml

@@ -61,7 +61,7 @@ tokio-tungstenite = { version = "0.20.1", features = ["native-tls"] }
 reqwest = { version = "0.11.20", features = ["stream"] }
 
 # Redis
-redis = { version = "0.26.0", features = ["default", "tokio-rustls-comp", "cluster-async", "connection-manager"] }
+redis = { version = "0.26.1", features = ["default", "tokio-rustls-comp", "cluster-async", "connection-manager"] }
 
 # Fluvio
 fluvio = {version = "0.23", features = ["openssl"]}

crates/arroyo-connectors/src/redis/mod.rs

@@ -122,7 +122,7 @@ async fn test_inner(
             .unwrap();
 
             redis::cmd("PING")
-                .query_async(&mut connection)
+                .query_async::<()>(&mut connection)
                 .await
                 .map_err(|e| anyhow!("Received error sending PING command: {:?}", e))?;
         }
@@ -138,7 +138,7 @@ async fn test_inner(
             .unwrap();
 
             redis::cmd("PING")
-                .query_async(&mut connection)
+                .query_async::<()>(&mut connection)
                 .await
                 .map_err(|e| anyhow!("Received error sending PING command: {:?}", e))?;
         }

crates/arroyo-formats/src/json/mod.rs

@@ -1,55 +1,10 @@
-use arrow::datatypes::{Field, Fields, SchemaRef};
-use arrow_array::builder::{ArrayBuilder, StringBuilder};
+use arrow::datatypes::{Field, Fields};
 use arrow_schema::DataType;
-use arroyo_rpc::formats::JsonFormat;
 use serde_json::{json, Value};
 use std::collections::HashMap;
 
 pub mod schema;
 
-pub fn deserialize_slice_json(
-    schema: &SchemaRef,
-    buffer: &mut [Box<dyn ArrayBuilder>],
-    format: &JsonFormat,
-    msg: &[u8],
-) -> Result<(), String> {
-    let msg = if format.confluent_schema_registry {
-        &msg[5..]
-    } else {
-        msg
-    };
-
-    if format.unstructured {
-        let (idx, _) = schema
-            .column_with_name("value")
-            .expect("no 'value' column for unstructured json format");
-        let array = buffer[idx]
-            .as_any_mut()
-            .downcast_mut::<StringBuilder>()
-            .expect("'value' column has incorrect type");
-
-        if format.include_schema {
-            // we need to deserialize it to pull out the payload
-            let v: Value = serde_json::from_slice(msg)
-                .map_err(|e| format!("Failed to deserialize json: {:?}", e))?;
-            let payload = v.get("payload").ok_or_else(|| {
-                "`include_schema` set to true, but record does not have a payload field".to_string()
-            })?;
-
-            array.append_value(serde_json::to_string(payload).unwrap());
-        } else {
-            array.append_value(
-                String::from_utf8(msg.to_vec()).map_err(|_| "data is not valid UTF-8")?,
-            );
-        };
-    } else {
-        serde_json::from_slice(msg)
-            .map_err(|e| format!("Failed to deserialize JSON into schema: {:?}", e))?;
-    }
-
-    Ok(())
-}
-
 pub fn field_to_json_schema(field: &Field) -> Value {
     match field.data_type() {
         arrow::datatypes::DataType::Null => {

docker/Dockerfile

@@ -8,9 +8,12 @@ ARG PROFILE=release
 
 RUN apt-get update && \
     apt-get -y install curl pkg-config unzip build-essential libssl-dev openssl \
-    cmake clang wget postgresql postgresql-client supervisor python3 python-is-python3 sudo bash libsasl2-dev && \
+    cmake clang wget postgresql postgresql-client supervisor sudo bash libsasl2-dev && \
     cargo install refinery_cli
 
+# Install Python (manually, as no bookworm package for 3.12)
+
+
 # Install node & pnpm
 RUN mkdir -p /etc/apt/keyrings && \
     curl -fsSL https://deb.nodesource.com/gpgkey/nodesource-repo.gpg.key | gpg --dearmor -o /etc/apt/keyrings/nodesource.gpg && \
@@ -57,6 +60,14 @@ RUN apt-get update && \
 COPY --from=builder /arroyo ./
 COPY --from=builder /usr/local/bin/protoc /usr/local/bin/protoc
 COPY --from=builder /app/include/google /usr/local/include/google
+COPY --from=builder /app/python /python
+
+RUN cp -r /python/bin/* /usr/local/bin/ && \
+    cp -r /python/include/* /usr/local/include/ && \
+    cp -r /python/lib/* /usr/local/lib/ && \
+    cp -r /python/share/* /usr/local/share/ && \
+    rm -rf /python && \
+    ldconfig
 
 ENV INSTALL_RUSTC=true \
     INSTALL_CLANG=true \
@@ -76,6 +87,15 @@ ENV ARROYO__API__RUN_HTTP_PORT=5115
 COPY --from=builder /arroyo ./
 COPY --from=builder /usr/local/bin/protoc /usr/local/bin/protoc
 COPY --from=builder /app/include/google /usr/local/include/google
+COPY --from=builder /app/python /python
+
+RUN cp -r /python/bin/* /usr/local/bin/ && \
+    cp -r /python/include/* /usr/local/include/ && \
+    cp -r /python/lib/* /usr/local/lib/ && \
+    cp -r /python/share/* /usr/local/share/ && \
+    rm -rf /python && \
+    ldconfig
+
 
 EXPOSE 5115
 ENTRYPOINT [ "/app/arroyo" ]

docker/install_deps.sh

@@ -7,14 +7,25 @@ echo env
 if [ "$(uname -m)" = "x86_64" ]; then
   export PROTO_ARCH="x86_64";
   export MOLD_ARCH="x86_64";
+  export PY_ARCH="x86_64";
 elif [ "$(uname -m)" = "aarch64" ]; then
   export PROTO_ARCH="aarch_64";
   export MOLD_ARCH="aarch64";
+  export PY_ARCH="aarch64";
 else
   echo "Unsupported architecture: $(uname -m)"
   exit 1;
 fi
 
+# Install Python 3.12 (manually, as no bookworm package for 3.12)
+curl -OL https://github.com/indygreg/python-build-standalone/releases/download/20240814/cpython-3.12.5+20240814-${PY_ARCH}-unknown-linux-gnu-install_only.tar.gz 
+tar xvfz cpython*.tar.gz
+cp -r python/bin/* /usr/local/bin/
+cp -r python/include/* /usr/local/include/
+cp -r python/lib/* /usr/local/lib/
+cp -r python/share/* /usr/local/share/
+ldconfig
+
 # Install mold
 curl -OL https://github.com/rui314/mold/releases/download/v1.11.0/mold-1.11.0-${MOLD_ARCH}-linux.tar.gz
 tar xvfz mold*.tar.gz