feat: determine bazel rule for an api path in generate (#14124)

Fixes googleapis/librarian#903

---------

Co-authored-by: Anthonios Partheniou <[email protected]>

Author: ohmayr

.generator/Dockerfile

@@ -22,6 +22,8 @@ RUN apt-get update && \
     # For downloading secure files
     wget \
     ca-certificates \
+    # For running bazelisk commands
+    openjdk-17-jdk \
     # --- Critical libraries for a complete Python build ---
     libssl-dev \
     zlib1g-dev \
@@ -35,6 +37,7 @@ RUN apt-get update && \
 
 # Set up environment variables for tool versions to make updates easier.
 ENV PYTHON_VERSION=3.11.5
+ENV BAZELISK_VERSION=v1.26.0
 
 # Create a symbolic link for `python3` to point to our specific version.
 ENV PATH /usr/local/bin/python3.11:$PATH
@@ -50,7 +53,9 @@ RUN wget https://www.python.org/ftp/python/${PYTHON_VERSION}/Python-${PYTHON_VER
 
 # TODO(https://github.com/googleapis/librarian/issues/904): Install protoc for gencode.
 
-# TODO(https://github.com/googleapis/librarian/issues/903): Install Bazelisk for building a client library.
+# Install Bazelisk
+RUN wget https://github.com/bazelbuild/bazelisk/releases/download/${BAZELISK_VERSION}/bazelisk-linux-amd64 -O /usr/local/bin/bazelisk && \
+    chmod +x /usr/local/bin/bazelisk
 
 # TODO(https://github.com/googleapis/librarian/issues/902): Create a dedicate non-root user and 
 # switch to the non-root user to run subsequent commands.

.generator/cli.py

@@ -17,11 +17,13 @@
 import logging
 import os
 import sys
+import subprocess
 
 logger = logging.getLogger()
 
 LIBRARIAN_DIR = "librarian"
 GENERATE_REQUEST_FILE = "generate-request.json"
+SOURCE_DIR = "source"
 
 
 def _read_json_file(path):
@@ -47,6 +49,39 @@ def handle_configure():
     logger.info("'configure' command executed.")
 
 
+def _determine_bazel_rule(api_path):
+    """Executes a `bazelisk query` to find a Bazel rule.
+
+    Args:
+        api_path (str): The API path to query for.
+
+    Returns:
+        str: The discovered Bazel rule.
+
+    Raises:
+        Exception: If the subprocess call fails or returns an empty result.
+    """
+    logger.info(f"Determining Bazel rule for api_path: '{api_path}'")
+    try:
+        query = f'filter("-py$", kind("rule", //{api_path}/...:*))'
+        command = ["bazelisk", "query", query]
+        result = subprocess.run(
+            command,
+            cwd=f"{SOURCE_DIR}/googleapis",
+            capture_output=True,
+            text=True,
+            check=True,
+        )
+        bazel_rule = result.stdout.strip()
+        if not bazel_rule:
+            raise ValueError(f"Bazelisk query `{query}` returned an empty bazel rule.")
+
+        logger.info(f"Found Bazel rule: {bazel_rule}")
+        return bazel_rule
+    except Exception as e:
+        raise ValueError(f"Bazelisk query `{query}` failed") from e
+
+
 def handle_generate():
     """The main coordinator for the code generation process.
 
@@ -61,12 +96,18 @@ def handle_generate():
     # Read a generate-request.json file
     try:
         request_data = _read_json_file(f"{LIBRARIAN_DIR}/{GENERATE_REQUEST_FILE}")
-    except Exception as e:
-        raise ValueError(
-            f"failed to read {LIBRARIAN_DIR}/{GENERATE_REQUEST_FILE}"
-        ) from e
+        library_id = request_data.get("id")
+        if not library_id:
+            raise ValueError("Request file is missing required 'id' field.")
+
+        for api in request_data.get("apis", []):
+            api_path = api.get("path")
+            if api_path:
+                bazel_rule = _determine_bazel_rule(api_path)
 
-    logger.info(json.dumps(request_data, indent=2))
+            logger.info(json.dumps(request_data, indent=2))
+    except Exception as e:
+        raise ValueError("Generation failed.") from e
 
     # TODO(https://github.com/googleapis/librarian/issues/448): Implement generate command and update docstring.
     logger.info("'generate' command executed.")

.generator/test_cli.py

@@ -16,11 +16,13 @@
 import pytest
 import json
 import logging
+import subprocess
 
-from unittest.mock import mock_open
+from unittest.mock import mock_open, MagicMock
 
 from cli import (
     _read_json_file,
+    _determine_bazel_rule,
     handle_generate,
     handle_build,
     handle_configure,
@@ -60,16 +62,51 @@ def test_handle_configure_success(caplog, mock_generate_request_file):
     assert "'configure' command executed." in caplog.text
 
 
-def test_handle_generate_success(caplog, mock_generate_request_file):
+def test_determine_bazel_rule_success(mocker, caplog):
+    """
+    Tests the happy path of _determine_bazel_rule.
+    """
+    caplog.set_level(logging.INFO)
+    mock_result = MagicMock(
+        stdout="//google/cloud/language/v1:google-cloud-language-v1-py\n"
+    )
+    mocker.patch("cli.subprocess.run", return_value=mock_result)
+
+    rule = _determine_bazel_rule("google/cloud/language/v1")
+
+    assert rule == "//google/cloud/language/v1:google-cloud-language-v1-py"
+    assert "Found Bazel rule" in caplog.text
+
+
+def test_determine_bazel_rule_command_fails(mocker, caplog):
+    """
+    Tests that an exception is raised if the subprocess command fails.
+    """
+    caplog.set_level(logging.INFO)
+    mocker.patch(
+        "cli.subprocess.run",
+        side_effect=subprocess.CalledProcessError(1, "cmd", stderr="Bazel error"),
+    )
+
+    with pytest.raises(ValueError):
+        _determine_bazel_rule("google/cloud/language/v1")
+
+    assert "Found Bazel rule" not in caplog.text
+
+
+def test_handle_generate_success(caplog, mock_generate_request_file, mocker):
     """
     Tests the successful execution path of handle_generate.
     """
     caplog.set_level(logging.INFO)
 
+    mock_determine_rule = mocker.patch(
+        "cli._determine_bazel_rule", return_value="mock-rule"
+    )
+
     handle_generate()
 
-    assert "google-cloud-language" in caplog.text
-    assert "'generate' command executed." in caplog.text
+    mock_determine_rule.assert_called_once_with("google/cloud/language/v1")
 
 
 def test_handle_generate_fail(caplog):