Feat: Use python-build-standalone over apt packages (#699)

- Use python-build-standalone as a drop-in replacement for custom images
without python
- Template the install script, python versions, and a few other runtime
variables

Resolve BE-1974

pkg/abstractions/image/build.go

@@ -1,14 +1,17 @@
 package image
 
 import (
+	"bytes"
 	"context"
 	"crypto/sha1"
 	_ "embed"
 	"encoding/hex"
 	"fmt"
 	"log"
 	"regexp"
+	"runtime"
 	"strings"
+	"text/template"
 	"time"
 
 	"github.com/google/uuid"
@@ -290,7 +293,11 @@ func (b *Builder) Build(ctx context.Context, opts *BuildOpts, outputChan chan co
 	checkPythonVersionCmd := fmt.Sprintf("%s --version", opts.PythonVersion)
 	if resp, err := client.Exec(containerId, checkPythonVersionCmd); (err != nil || !resp.Ok) && !micromambaEnv {
 		outputChan <- common.OutputMsg{Done: false, Success: false, Msg: fmt.Sprintf("%s not detected, installing it for you...\n", opts.PythonVersion)}
-		installCmd := getPythonInstallCommand(opts.PythonVersion)
+		installCmd, err := getPythonStandaloneInstallCommand(b.config.ImageService.Runner.PythonStandalone, opts.PythonVersion)
+		if err != nil {
+			outputChan <- common.OutputMsg{Done: true, Success: false, Msg: err.Error() + "\n"}
+			return err
+		}
 		opts.Commands = append([]string{installCmd}, opts.Commands...)
 	}
 
@@ -493,6 +500,55 @@ func getPythonInstallCommand(pythonVersion string) string {
 	return fmt.Sprintf("%s && add-apt-repository ppa:deadsnakes/ppa && apt-get update && apt-get install -q -y %s && %s", baseCmd, installCmd, postInstallCmd)
 }
 
+// PythonStandaloneTemplate is used to render the standalone python install script
+type PythonStandaloneTemplate struct {
+	PythonVersion string
+
+	// Architecture, OS, and Vendor are determined at runtime
+	Architecture string
+	OS           string
+	Vendor       string
+}
+
+func getPythonStandaloneInstallCommand(config types.PythonStandaloneConfig, pythonVersion string) (string, error) {
+	var arch string
+	switch runtime.GOARCH {
+	case "amd64":
+		arch = "x86_64"
+	case "arm64":
+		arch = "aarch64"
+	default:
+		return "", errors.New("unsupported architecture for python standalone install")
+	}
+
+	var vendor, os string
+	switch runtime.GOOS {
+	case "linux":
+		vendor, os = "unknown", "linux"
+	case "darwin":
+		vendor, os = "apple", "darwin"
+	default:
+		return "", errors.New("unsupported OS for python standalone install")
+	}
+
+	tmpl, err := template.New("standalonePython").Parse(config.InstallScriptTemplate)
+	if err != nil {
+		return "", err
+	}
+
+	var output bytes.Buffer
+	if err := tmpl.Execute(&output, PythonStandaloneTemplate{
+		PythonVersion: config.Versions[pythonVersion],
+		Architecture:  arch,
+		OS:            os,
+		Vendor:        vendor,
+	}); err != nil {
+		return "", err
+	}
+
+	return output.String(), nil
+}
+
 func generatePipInstallCommand(pythonPackages []string, pythonVersion string) string {
 	flagLines, packages := parseFlagLinesAndPackages(pythonPackages)
 

pkg/common/config.default.yaml

@@ -78,6 +78,25 @@ imageService:
       micromamba3.10: micromamba3.10-latest
       micromamba3.11: micromamba3.11-latest
       micromamba3.12: micromamba3.12-latest
+    pythonStandalone:
+      versions:
+        python3.8: 3.8.20
+        python3.9: 3.9.20
+        python3.10: 3.10.5
+        python3.11: 3.11.10
+        python3.12: 3.12.7
+      installScriptTemplate: |
+        apt-get update -q && \
+        apt-get install -q -y build-essential curl git && \
+        curl -fsSL -o python.tgz 'https://github.com/indygreg/python-build-standalone/releases/download/20241016/cpython-{{.PythonVersion}}+20241016-{{.Architecture}}-{{.Vendor}}-{{.OS}}-gnu-install_only.tar.gz' && \
+        tar -xzf python.tgz -C /usr/local --strip-components 1 && \
+        rm -f python.tgz && \
+        rm -f /usr/bin/python && \
+        rm -f /usr/bin/python3 && \
+        ln -s /usr/local/bin/python3 /usr/bin/python && \
+        ln -s /usr/local/bin/python3 /usr/bin/python3 && \
+        ln -s /usr/local/bin/pip3 /usr/bin/pip && \
+        ln -s /usr/local/bin/pip3 /usr/bin/pip3
 worker:
   pools:
     default:
@@ -204,4 +223,4 @@ abstractions:
       Do not return the same marker data back to the user in consecutive responses.
 
     stepIntervalS: 1
-    sessionInactivityTimeoutS: 10
+    sessionInactivityTimeoutS: 10

pkg/types/config.go

@@ -132,9 +132,15 @@ type S3ImageRegistryConfig struct {
 }
 
 type RunnerConfig struct {
-	BaseImageName     string            `key:"baseImageName" json:"base_image_name"`
-	BaseImageRegistry string            `key:"baseImageRegistry" json:"base_image_registry"`
-	Tags              map[string]string `key:"tags" json:"tags"`
+	BaseImageName     string                 `key:"baseImageName" json:"base_image_name"`
+	BaseImageRegistry string                 `key:"baseImageRegistry" json:"base_image_registry"`
+	Tags              map[string]string      `key:"tags" json:"tags"`
+	PythonStandalone  PythonStandaloneConfig `key:"pythonStandalone" json:"python_standalone"`
+}
+
+type PythonStandaloneConfig struct {
+	Versions              map[string]string `key:"versions" json:"versions"`
+	InstallScriptTemplate string            `key:"installScriptTemplate" json:"install_script_template"`
 }
 
 type StorageConfig struct {