added and propagated cli option to skip extraction

onekey-sec · Oct 12, 2022 · f5a9490 · f5a9490
1 parent 91a2b04
commit f5a9490
Show file tree

Hide file tree

Showing 2 changed files with 22 additions and 10 deletions.
diff --git a/unblob/cli.py b/unblob/cli.py
@@ -151,6 +151,14 @@ def __init__(
     type=click.Path(path_type=Path),
     help="File to store metadata generated during the extraction process (in JSON format).",
 )
+@click.option(
+    "-s",
+    "--skip_extraction",
+    "skip_extraction",
+    is_flag=True,
+    show_default=True,
+    help="Only carve chunks and skip further extraction",
+)
 @click.option(
     "-k",
     "--keep-extracted-chunks",
@@ -176,6 +184,7 @@ def cli(
     depth: int,
     entropy_depth: int,
     skip_magic: Iterable[str],
+    skip_extraction: bool,
     keep_extracted_chunks: bool,
     handlers: Handlers,
     plugins_path: Optional[Path],
@@ -194,6 +203,7 @@ def cli(
         max_depth=depth,
         entropy_depth=entropy_depth,
         entropy_plot=bool(verbose >= 3),
+        skip_extraction=skip_extraction,
         skip_magic=skip_magic,
         process_num=process_num,
         handlers=handlers,

diff --git a/unblob/processing.py b/unblob/processing.py
@@ -72,6 +72,7 @@ class ExtractionConfig:
     entropy_plot: bool = False
     max_depth: int = DEFAULT_DEPTH
     skip_magic: Iterable[str] = DEFAULT_SKIP_MAGIC
+    skip_extraction: bool = False
     process_num: int = DEFAULT_PROCESS_NUM
     keep_extracted_chunks: bool = False
     extract_suffix: str = "_extract"
@@ -367,19 +368,20 @@ def _extract_chunk(self, file, chunk: ValidChunk):
             carved_path = inpath
 
         extraction_reports = []
-        try:
-            chunk.extract(inpath, extract_dir)
+        if not self.config.skip_extraction:
+            try:
+                chunk.extract(inpath, extract_dir)
 
-            if carved_path and not self.config.keep_extracted_chunks:
-                logger.debug("Removing extracted chunk", path=carved_path)
-                carved_path.unlink()
+                if carved_path and not self.config.keep_extracted_chunks:
+                    logger.debug("Removing extracted chunk", path=carved_path)
+                    carved_path.unlink()
 
-        except ExtractError as e:
-            extraction_reports.extend(e.reports)
+            except ExtractError as e:
+                extraction_reports.extend(e.reports)
 
-        except Exception as exc:
-            logger.exception("Unknown error happened while extracting chunk")
-            extraction_reports.append(UnknownError(exception=exc))
+            except Exception as exc:
+                logger.exception("Unknown error happened while extracting chunk")
+                extraction_reports.append(UnknownError(exception=exc))
 
         self.result.add_report(chunk.as_report(extraction_reports))