Add envsubst.py script (#3425)

## Changes
Use the python script instead of the installed `envsubst` binary in
acceptance tests.

## Why
envsubst is not installed by default on Ubuntu and has to be installed
by installing the GNU gettext tools. Since we will start running
acceptance tests on DBR soon, it's better to use a python script we own
instead since that means one less dependency to install when running
tests on DBR.

I tried installing envsubst on DBR, but apt installations are not
allowed on serverless. It's easier to have a simple python script
instead:
```
Failed to install envsubst: Command '['apt-get', 'update']' returned non-zero exit status 100.
stdout: Reading package lists...

stderr: E: Could not open lock file /var/lib/apt/lists/lock - open (13: Permission denied)
E: Unable to lock directory /var/lib/apt/lists/
```

## Tests
Existing tests pass.

Author: shreyas-goenka

.vscode/settings.json

@@ -19,6 +19,8 @@
     "jupyter.interactiveWindow.cellMarker.codeRegex": "^# COMMAND ----------|^# Databricks notebook source|^(#\\s*%%|#\\s*\\<codecell\\>|#\\s*In\\[\\d*?\\]|#\\s*In\\[ \\])",
     "jupyter.interactiveWindow.cellMarker.default": "# COMMAND ----------",
     "files.associations": {
-        "script": "shellscript"
+        "script": "shellscript",
+        "script.prepare": "shellscript",
+        "script.cleanup": "shellscript"
     }
 }

acceptance/bin/envsubst.py

@@ -0,0 +1,51 @@
+#!/usr/bin/env python3
+"""
+This script implements functionality equivalent to envsubst. We use a python
+script instead because that way we do not need to install it when we run integration
+tests on DBR.
+
+Usage: envsubst.py
+
+Substitutes environment variables in shell format strings.
+Reads from stdin and writes to stdout.
+
+Examples:
+  echo 'Hello $USER' | python envsubst.py
+  echo 'Hello $USER from $HOME' | python envsubst.py
+"""
+
+import sys
+import os
+import re
+
+
+def substitute_variables(text):
+    """
+    Substitute environment variables in text.
+
+    Args:
+        text: Input text containing variable references
+
+    Returns:
+        Text with variables substituted
+    """
+
+    def replace_var(match):
+        var_name = match.group(1) or match.group(2)
+
+        # If the environment variable is not set, replace with an empty string.
+        return os.environ.get(var_name, "")
+
+    # Match both $VAR and ${VAR} formats
+    pattern = r"\$\{([A-Za-z_][A-Za-z0-9_]*)\}|\$([A-Za-z_][A-Za-z0-9_]*)"
+    return re.sub(pattern, replace_var, text)
+
+
+def main():
+    input_text = sys.stdin.read()
+    output_text = substitute_variables(input_text)
+    sys.stdout.write(output_text)
+
+
+if __name__ == "__main__":
+    main()

acceptance/bundle/validate/empty_resources/script.py

@@ -1,4 +1,5 @@
 import os
+import sys
 
 SECTIONS = [
     "jobs",

acceptance/script.prepare

@@ -86,3 +86,10 @@ venv_activate() {
         source .venv/bin/activate
     fi
 }
+
+envsubst() {
+    # We need to disable MSYS_NO_PATHCONV when running the python script.
+    # This is because the python interpreter is otherwise unable to find the python script
+    # when MSYS_NO_PATHCONV is enabled.
+    env -u MSYS_NO_PATHCONV envsubst.py
+}