Add optional JSON logging

pyproject.toml

@@ -13,6 +13,7 @@ dependencies = [
     "isort~=5.12.0",
     "libcst~=1.1.0",
     "pylint~=3.0.0",
+    "python-json-logger~=2.0.0",
     "PyYAML~=6.0.0",
     "semgrep~=1.45.0",
     "wrapt~=1.15.0",

src/codemodder/cli.py

@@ -139,9 +139,13 @@ def parse_args(argv, codemod_registry):
         "--log-format",
         type=OutputFormat,
         default=OutputFormat.HUMAN,
-        choices=[str(x).split(".")[-1].lower() for x in list(OutputFormat)],
+        choices=[OutputFormat.HUMAN, OutputFormat.JSON],
         help="the format for the log output",
     )
+    parser.add_argument(
+        "--project-name",
+        help="optional descriptive name for the project used in log output",
+    )
     parser.add_argument(
         "--path-exclude",
         action=CsvListAction,

src/codemodder/codemodder.py

@@ -129,7 +129,7 @@ def run(original_args) -> int:
         )
         return 1
 
-    configure_logger(argv.verbose)
+    configure_logger(argv.verbose, argv.log_format, argv.project_name)
 
     log_section("startup")
     logger.info("codemodder: python/%s", __VERSION__)

src/codemodder/logging.py

@@ -1,6 +1,9 @@
 from enum import Enum
 import logging
 import sys
+from typing import Optional
+
+from pythonjsonlogger import jsonlogger
 
 logger = logging.getLogger("codemodder")
 
@@ -13,6 +16,28 @@
     HUMAN = "human"
     JSON = "json"
 
+    def __str__(self):
+        """For rendering properly in argparse help."""
+        return self.value.lower()
+
+
+class CodemodderJsonFormatter(jsonlogger.JsonFormatter):
+    project_name: Optional[str]
+
+    def __init__(self, *args, project_name: Optional[str] = None, **kwargs):
+        self.project_name = project_name
+        super().__init__(*args, **kwargs)
+
+    def add_fields(self, log_record, record, message_dict):
+        super().add_fields(log_record, record, message_dict)
+        log_record["timestamp"] = log_record.pop("asctime")
+        log_record.move_to_end("timestamp", last=False)
+        log_record["level"] = record.levelname.upper()
+        log_record["file"] = record.filename
+        log_record["line"] = record.lineno
+        if self.project_name:
+            log_record["project-name"] = self.project_name
+
 
 def log_section(section_name: str):
     """
@@ -30,22 +55,33 @@
         logger.log(level, "  - %s", predicate(item) if predicate else item)
 
 
-def configure_logger(verbose: bool):
+def configure_logger(
+    verbose: bool, log_format: OutputFormat, project_name: Optional[str] = None
+):
     """
     Configure the logger based on the verbosity level.
     """
     log_level = logging.DEBUG if verbose else logging.INFO
 
-    # TODO: this should all be conditional on the output format
     stdout_handler = logging.StreamHandler(sys.stdout)
     stdout_handler.setLevel(log_level)
-    stdout_handler.addFilter(lambda record: record.levelno <= logging.WARNING)
+    handlers = [stdout_handler]
 
-    stderr_handler = logging.StreamHandler(sys.stderr)
-    stderr_handler.setLevel(logging.ERROR)
+    match log_format:
+        case OutputFormat.HUMAN:
+            stdout_handler.addFilter(lambda record: record.levelno <= logging.WARNING)
+            stderr_handler = logging.StreamHandler(sys.stderr)
+            stderr_handler.setLevel(logging.ERROR)
+            handlers.append(stderr_handler)
+        case OutputFormat.JSON:
+            formatter = CodemodderJsonFormatter(
+                "%(asctime) %(level) %(message) %(file) %(line)",
+                project_name=project_name,
+            )
+            stdout_handler.setFormatter(formatter)
 
     logging.basicConfig(
         format="%(message)s",
         level=log_level,
-        handlers=[stdout_handler, stderr_handler],
+        handlers=handlers,
     )

tests/test_logging.py

@@ -0,0 +1,18 @@
+from pythonjsonlogger import jsonlogger
+
+from codemodder.logging import OutputFormat, configure_logger
+
+
+def test_json_logger(mocker):
+    basic_config = mocker.patch("logging.basicConfig")
+    configure_logger(False, OutputFormat.JSON, "test-project")
+    assert basic_config.call_count == 1
+    assert basic_config.call_args[1]["format"] == "%(message)s"
+    assert isinstance(
+        basic_config.call_args[1]["handlers"][0].formatter,
+        jsonlogger.JsonFormatter,
+    )
+    assert (
+        basic_config.call_args[1]["handlers"][0].formatter.project_name
+        == "test-project"
+    )