From 6d291a02f43e0c1dc461074c512b9597589db577 Mon Sep 17 00:00:00 2001
From: Keshav Priyadarshi <git@keshav.space>
Date: Wed, 31 Jan 2024 21:25:03 +0530
Subject: [PATCH] Support extraction of JavaScript map files

fixes https://github.com/nexB/scancode-toolkit/issues/3637

Signed-off-by: Keshav Priyadarshi <git@keshav.space>
---
 src/extractcode/__init__.py   |  4 ++
 src/extractcode/archive.py    | 22 ++++++++++
 src/extractcode/source_map.py | 78 +++++++++++++++++++++++++++++++++++
 3 files changed, 104 insertions(+)
 create mode 100644 src/extractcode/source_map.py

diff --git a/src/extractcode/__init__.py b/src/extractcode/__init__.py
index fb6095d..efe2d88 100644
--- a/src/extractcode/__init__.py
+++ b/src/extractcode/__init__.py
@@ -42,6 +42,7 @@
 file_system = 5
 patches = 6
 special_package = 7
+sources_map = 8
 
 kind_labels = {
     1: 'docs',
@@ -51,6 +52,7 @@
     5: 'file_system',
     6: 'patches',
     7: 'special_package',
+    8: 'sources_map',
 }
 
 # note: we do not include special_package in all_kinds by default
@@ -62,6 +64,7 @@
     docs,
     patches,
     special_package,
+    sources_map
 )
 
 default_kinds = (
@@ -79,6 +82,7 @@
     'doc': (docs,),
     'patch': (patches,),
     'special_package': (special_package,),
+    'sources_map': (sources_map,),
 }
 
 
diff --git a/src/extractcode/archive.py b/src/extractcode/archive.py
index d52399e..1ad180b 100644
--- a/src/extractcode/archive.py
+++ b/src/extractcode/archive.py
@@ -24,10 +24,12 @@
 from extractcode import regular_nested
 from extractcode import file_system
 from extractcode import patches
+from extractcode import sources_map
 from extractcode import special_package
 
 from extractcode import libarchive2
 from extractcode import patch
+from extractcode import source_map
 from extractcode import sevenzip
 from extractcode import vmimage
 
@@ -506,6 +508,7 @@ def try_to_extract(location, target_dir, extractor):
 extract_ishield = sevenzip.extract
 extract_Z = sevenzip.extract
 extract_xarpkg = sevenzip.extract
+extract_source_map = source_map.extract
 
 # Archive handlers.
 ####################
@@ -1133,6 +1136,24 @@ def try_to_extract(location, target_dir, extractor):
     strict=True
 )
 
+SourceMapFileHandler = Handler(
+    name='Source Map File',
+    filetypes=('json data',),
+    mimetypes=('application/json',),
+    extensions=(
+        '.js.map',
+        '.ts.map',
+        '.css.map',
+        '.less.map',
+        '.scss.map',
+        '.soy.map',
+        '.jsx.map',
+        ),
+    kind=sources_map,
+    extractors=[extract_source_map],
+    strict=True,
+)
+
 # Actual list of handlers
 
 archive_handlers = [
@@ -1194,6 +1215,7 @@ def try_to_extract(location, target_dir, extractor):
     QCOWHandler,
     VMDKHandler,
     VirtualBoxHandler,
+    SourceMapFileHandler,
 ]
 
 # only support extracting patches if patch is installed. This is not a default
diff --git a/src/extractcode/source_map.py b/src/extractcode/source_map.py
new file mode 100644
index 0000000..d3f46a8
--- /dev/null
+++ b/src/extractcode/source_map.py
@@ -0,0 +1,78 @@
+#
+# Copyright (c) nexB Inc. and others. All rights reserved.
+# ScanCode is a trademark of nexB Inc.
+# SPDX-License-Identifier: Apache-2.0
+# See http://www.apache.org/licenses/LICENSE-2.0 for the license text.
+# See https://github.com/nexB/extractcode for support or download.
+# See https://aboutcode.org for more information about nexB OSS projects.
+#
+
+import json
+import os.path
+import posixpath
+
+from commoncode import fileutils
+from commoncode import paths
+
+import extractcode
+
+"""
+Utilities to parse source map files and treat them as if they were
+archives containing files.
+"""
+
+
+def extract(location, target_dir):
+    """
+    Extract each source in sourcesContent list of a map file at `location` as
+    files in a target_dir directory tree mimicking the directory in which the
+    sources would be present.
+    
+    Return a list of warning messages. Raise Exception errors.
+    """
+    for path, content in extract_source_content_from_map(location):
+        # Convert path to safe posix path
+        map_subfile_path = paths.safe_path(path, preserve_spaces=True)
+
+        # Create directories
+        parent_dir = posixpath.dirname(map_subfile_path)
+        parent_target_dir = os.path.join(target_dir, parent_dir)
+        fileutils.create_dir(parent_target_dir)
+
+        subfile_path = os.path.join(target_dir, map_subfile_path)
+        with open(subfile_path, "w") as subfile:
+            subfile.write(content)
+
+        return []
+
+
+def extract_source_content_from_map(location):
+    """
+    Return a list of tuples of (source, content)
+    for each source in sourcesContent of a map file at location.
+
+    Raise an exception if the file is not a JSON file or cannot be parsed.
+    """
+    try:
+        with open(location, "r") as map_file:
+            map_data = json.load(map_file)
+    except json.JSONDecodeError as e:
+        msg = f"Unable to decode map file:{location} {e}"
+        raise extractcode.ExtractErrorFailedToExtract(msg)
+
+    if "sourcesContent" in map_data:
+        sources_content = map_data["sourcesContent"]
+        sources = map_data.get("sources", [])
+
+        # Inconsistent source map. In a valid source map, each entry in the ``sources``
+        # list should have a corresponding entry in the ``sourcesContent`` list.
+        # Use dummy filenames as `source` path in such scenario.
+        if len(sources) != len(sources_content):
+            sources = [
+                f"source_content{i + 1}.txt" for i in range(len(sources_content))
+            ]
+
+        sources_and_content = list(zip(sources, sources_content))
+        return sources_and_content
+
+    return []