fix: local connector output filename when a single file is being processed (#879)

* fix string processing error for _output_filename

* Add docstring and type hint, update CHANGELOG, update version

* update test fixture

* simple code change commit to retrigger ci checks

* update test fixture - after brew install tesseract-lang

* Update ingest test fixtures (#882)

Co-authored-by: ahmetmeleq <[email protected]>

* correct CHANGELOG

* correct CHANGELOG

---------

Co-authored-by: Unstructured-DevOps <[email protected]>
Co-authored-by: ahmetmeleq <[email protected]>

Author: ahmetmeleq

.pre-commit-config.yaml

@@ -8,6 +8,7 @@ repos:
       - id: check-json
       - id: check-xml
       - id: end-of-file-fixer
+        exclude: \.json$
         include: \.py$
       - id: trailing-whitespace
       - id: mixed-line-ending

CHANGELOG.md

@@ -1,4 +1,4 @@
-## 0.8.0-dev0
+## 0.8.0-dev1
 
 ### Enhancements
 
@@ -8,6 +8,7 @@
 
 ### Fixes
 * Fix KeyError when `isd_to_elements` doesn't find a type
+* Fix _output_filename for local connector, allowing single files to be written correctly to the disk
 
 ### BREAKING CHANGES
 

…example-docs/english-and-korean.png.json …-single-file/english-and-korean.png.jsontest_unstructured_ingest/expected-structured-output/local-single-file/example-docs/english-and-korean.png.json renamed to test_unstructured_ingest/expected-structured-output/local-single-file/english-and-korean.png.json test_unstructured_ingest/expected-structured-output/local-single-file/example-docs/english-and-korean.png.json renamed to test_unstructured_ingest/expected-structured-output/local-single-file/english-and-korean.png.json

@@ -1 +1 @@
-__version__ = "0.8.0-dev0"  # pragma: no cover
+__version__ = "0.8.0-dev1"  # pragma: no cover

unstructured/__version__.py

@@ -51,11 +51,18 @@ def get_file(self):
         pass
 
     @property
-    def _output_filename(self):
-        return (
-            Path(self.standard_config.output_dir)
-            / f"{self.path.replace(f'{self.config.input_path}/', '')}.json"
+    def _output_filename(self) -> Path:
+        """Returns output filename for the doc
+        If input path argument is a file itself, it returns the filename of the doc.
+        If input path argument is a folder, it returns the relative path of the doc.
+        """
+        input_path = Path(self.config.input_path)
+        basename = (
+            f"{Path(self.path).name}.json"
+            if input_path.is_file()
+            else f"{Path(self.path).relative_to(input_path)}.json"
         )
+        return Path(self.standard_config.output_dir) / basename
 
 
 class LocalConnector(BaseConnector):