Changes os.path to pathlib.Path usage

changed args.rules , args.signatures types in handle_common_args.

tests/fixtures.py

@@ -14,6 +14,7 @@ import itertools
 import contextlib
 import collections
 from typing import Set, Dict
+from pathlib import Path
 from functools import lru_cache
 
 import pytest
@@ -44,9 +45,9 @@ from capa.features.address import Address
 from capa.features.extractors.base_extractor import BBHandle, InsnHandle, FunctionHandle
 from capa.features.extractors.dnfile.extractor import DnfileFeatureExtractor
 
-CD = os.path.dirname(__file__)
-DOTNET_DIR = os.path.join(CD, "data", "dotnet")
-DNFILE_TESTFILES = os.path.join(DOTNET_DIR, "dnfile-testfiles")
+CD = Path(__file__).resolve().parent
+DOTNET_DIR = CD / "data" / "dotnet"
+DNFILE_TESTFILES = DOTNET_DIR / "dnfile-testfiles"
 
 
 @contextlib.contextmanager
@@ -94,11 +95,11 @@ def get_viv_extractor(path):
     import capa.features.extractors.viv.extractor
 
     sigpaths = [
-        os.path.join(CD, "data", "sigs", "test_aulldiv.pat"),
-        os.path.join(CD, "data", "sigs", "test_aullrem.pat.gz"),
-        os.path.join(CD, "..", "sigs", "1_flare_msvc_rtf_32_64.sig"),
-        os.path.join(CD, "..", "sigs", "2_flare_msvc_atlmfc_32_64.sig"),
-        os.path.join(CD, "..", "sigs", "3_flare_common_libs.sig"),
+        CD / "data" / "sigs" / "test_aulldiv.pat",
+        CD / "data" / "sigs" / "test_aullrem.pat.gz",
+        CD.parent / "sigs" / "1_flare_msvc_rtf_32_64.sig",
+        CD.parent / "sigs" / "2_flare_msvc_atlmfc_32_64.sig",
+        CD.parent / "sigs" / "3_flare_common_libs.sig",
     ]
 
     if "raw32" in path:

tests/test_binja_features.py

@@ -6,6 +6,7 @@
 #  is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and limitations under the License.
 import logging
+from pathlib import Path
 
 import fixtures
 from fixtures import *
@@ -52,8 +53,8 @@ def test_binja_feature_counts(sample, scope, feature, expected):
 
 @pytest.mark.skipif(binja_present is False, reason="Skip binja tests if the binaryninja Python API is not installed")
 def test_standalone_binja_backend():
-    CD = os.path.dirname(__file__)
-    test_path = os.path.join(CD, "..", "tests", "data", "Practical Malware Analysis Lab 01-01.exe_")
+    CD = Path(__file__).resolve().parent
+    test_path = CD / ".." / "tests" / "data" / "Practical Malware Analysis Lab 01-01.exe_"
     assert capa.main.main([test_path, "-b", capa.main.BACKEND_BINJA]) == 0
 
 

tests/test_ida_features.py

@@ -46,14 +46,14 @@ import io
 import sys
 import inspect
 import logging
-import os.path
 import binascii
 import traceback
+from pathlib import Path
 
 import pytest
 
 try:
-    sys.path.append(os.path.dirname(__file__))
+    sys.path.append(str(Path(__file__).parent))
     import fixtures
     from fixtures import *
 finally:

tests/test_proto.py

@@ -144,7 +144,7 @@ def assert_meta(meta: rd.Metadata, dst: capa_pb2.Metadata):
     assert meta.analysis.arch == dst.analysis.arch
     assert meta.analysis.os == dst.analysis.os
     assert meta.analysis.extractor == dst.analysis.extractor
-    assert list(meta.analysis.rules) == dst.analysis.rules
+    assert list(str(r) for r in meta.analysis.rules) == dst.analysis.rules
     assert capa.render.proto.addr_to_pb2(meta.analysis.base_address) == dst.analysis.base_address
 
     assert len(meta.analysis.layout.functions) == len(dst.analysis.layout.functions)

tests/test_rule_cache.py

@@ -76,12 +76,12 @@ def test_ruleset_cache_save_load():
 
     path = capa.rules.cache.get_cache_path(cache_dir, id)
     try:
-        os.remove(path)
+        path.unlink()
     except OSError:
         pass
 
     capa.rules.cache.cache_ruleset(cache_dir, rs)
-    assert os.path.exists(path)
+    assert path.exists()
 
     assert capa.rules.cache.load_cached_ruleset(cache_dir, content) is not None
 
@@ -93,23 +93,23 @@ def test_ruleset_cache_invalid():
     cache_dir = capa.rules.cache.get_default_cache_directory()
     path = capa.rules.cache.get_cache_path(cache_dir, id)
     try:
-        os.remove(path)
+        path.unlink()
     except OSError:
         pass
 
     capa.rules.cache.cache_ruleset(cache_dir, rs)
-    assert os.path.exists(path)
+    assert path.exists()
 
-    with open(path, "rb") as f:
-        buf = f.read()
+    buf = path.read_bytes()
 
-    # corrupt the magic header
+    # Corrupt the magic header
     buf = b"x" + buf[1:]
 
-    with open(path, "wb") as f:
-        f.write(buf)
+    # Write the modified contents back to the file
+    path.write_bytes(buf)
 
-    assert os.path.exists(path)
+    # Check if the file still exists
+    assert path.exists()
     assert capa.rules.cache.load_cached_ruleset(cache_dir, content) is None
     # the invalid cache should be deleted
-    assert not os.path.exists(path)
+    assert not path.exists()

tests/test_scripts.py

@@ -10,27 +10,28 @@ import os
 import sys
 import textwrap
 import subprocess
+from pathlib import Path
 
 import pytest
 from fixtures import *
 
-CD = os.path.dirname(__file__)
+CD = Path(__file__).resolve().parent
 
 
 def get_script_path(s):
-    return os.path.join(CD, "..", "scripts", s)
+    return CD / ".." / "scripts" / s
 
 
 def get_file_path():
-    return os.path.join(CD, "data", "9324d1a8ae37a36ae560c37448c9705a.exe_")
+    return CD / "data" / "9324d1a8ae37a36ae560c37448c9705a.exe_"
 
 
 def get_rules_path():
-    return os.path.join(CD, "..", "rules")
+    return CD / ".." / "rules"
 
 
 def get_rule_path():
-    return os.path.join(get_rules_path(), "lib", "allocate-memory.yml")
+    return get_rules_path() / "lib" / "allocate-memory.yml"
 
 
 @pytest.mark.parametrize(
@@ -52,13 +53,17 @@ def test_scripts(script, args):
     assert p.returncode == 0
 
 
-def test_bulk_process(tmpdir):
+def test_bulk_process(tmp_path):
     # create test directory to recursively analyze
-    t = tmpdir.mkdir("test")
-    with open(os.path.join(CD, "data", "ping_täst.exe_"), "rb") as f:
-        t.join("test.exe_").write_binary(f.read())
+    t = tmp_path / "test"
+    t.mkdir()
 
-    p = run_program(get_script_path("bulk-process.py"), [t.dirname])
+    source_file = Path(__file__).resolve().parent / "data" / "ping_täst.exe_"
+    dest_file = t / "test.exe_"
+
+    dest_file.write_bytes(source_file.read_bytes())
+
+    p = run_program(get_script_path("bulk-process.py"), [t.parent])
     assert p.returncode == 0
 
 
@@ -68,19 +73,18 @@ def run_program(script_path, args):
     return subprocess.run(args, stdout=subprocess.PIPE)
 
 
-def test_proto_conversion(tmpdir):
-    t = tmpdir.mkdir("proto-test")
+def test_proto_conversion(tmp_path):
+    t = tmp_path / "proto-test"
+    t.mkdir()
+    json_file = Path(__file__).resolve().parent / "data" / "rd" / "Practical Malware Analysis Lab 01-01.dll_.json"
 
-    json = os.path.join(CD, "data", "rd", "Practical Malware Analysis Lab 01-01.dll_.json")
-
-    p = run_program(get_script_path("proto-from-results.py"), [json])
+    p = run_program(get_script_path("proto-from-results.py"), [json_file])
     assert p.returncode == 0
 
-    pb = os.path.join(t, "pma.pb")
-    with open(pb, "wb") as f:
-        f.write(p.stdout)
+    pb_file = t / "pma.pb"
+    pb_file.write_bytes(p.stdout)
 
-    p = run_program(get_script_path("proto-to-results.py"), [pb])
+    p = run_program(get_script_path("proto-to-results.py"), [pb_file])
     assert p.returncode == 0
 
     assert p.stdout.startswith(b'{\n  "meta": ') or p.stdout.startswith(b'{\r\n  "meta": ')