From ea4fdd913483887b6d234da978a1adbcce0fd5c0 Mon Sep 17 00:00:00 2001
From: Debanjum Singh Solanky <debanjum@gmail.com>
Date: Sun, 21 Aug 2022 19:41:40 +0300
Subject: [PATCH] Fix logic to ignore notes with no body. Add tests to prevent
 regression

- Notes with empty newlines in body were not being ignored
- Add regression tests to avoid above regression in org_to_jsonl conversion
---
 src/processor/org_mode/org_to_jsonl.py |  5 ++-
 tests/test_org_to_jsonl.py             | 62 ++++++++++++++++++++++++++
 2 files changed, 65 insertions(+), 2 deletions(-)
 create mode 100644 tests/test_org_to_jsonl.py

diff --git a/src/processor/org_mode/org_to_jsonl.py b/src/processor/org_mode/org_to_jsonl.py
index caad7715..ea2962f5 100644
--- a/src/processor/org_mode/org_to_jsonl.py
+++ b/src/processor/org_mode/org_to_jsonl.py
@@ -1,6 +1,7 @@
 #!/usr/bin/env python3
 
 # Standard Packages
+import re
 import json
 import argparse
 import pathlib
@@ -71,14 +72,14 @@ def extract_org_entries(org_files):
     return entries
 
 
-def convert_org_entries_to_jsonl(entries, verbose=0):
+def convert_org_entries_to_jsonl(entries, verbose=0) -> str:
     "Convert each Org-Mode entries to JSON and collate as JSONL"
     jsonl = ''
     for entry in entries:
         entry_dict = dict()
 
         # Ignore title notes i.e notes with just headings and empty body
-        if not entry.Body() or entry.Body().strip(empty_escape_sequences) == "":
+        if not entry.Body() or re.sub(r'\n|\t|\r| ', '', entry.Body()) == "":
             continue
 
         entry_dict["compiled"] = f'{entry.Heading()}.'
diff --git a/tests/test_org_to_jsonl.py b/tests/test_org_to_jsonl.py
new file mode 100644
index 00000000..cadd4a6a
--- /dev/null
+++ b/tests/test_org_to_jsonl.py
@@ -0,0 +1,62 @@
+# Standard Packages
+import json
+from posixpath import split
+
+# Internal Packages
+from src.processor.org_mode.org_to_jsonl import convert_org_entries_to_jsonl, extract_org_entries
+from src.utils.helpers import is_none_or_empty
+
+
+def test_entry_with_empty_body_line_to_jsonl(tmp_path):
+    '''Ensure entries with empty body are ignored.
+    Property drawers not considered Body. Ignore control characters for evaluating if Body empty.'''
+    # Arrange
+    entry = f'''*** Heading
+    :PROPERTIES:
+    :ID:       42-42-42
+    :END:
+    \t\r\n 
+    '''
+    orgfile = create_file(tmp_path, entry)
+
+    # Act
+    # Extract Entries from specified Org files
+    entries = extract_org_entries(org_files=[orgfile])
+
+    # Process Each Entry from All Notes Files
+    jsonl_data = convert_org_entries_to_jsonl(entries)
+
+    # Assert
+    assert is_none_or_empty(jsonl_data)
+
+
+def test_entry_with_body_to_jsonl(tmp_path):
+    "Ensure entries with valid body text are loaded."
+    # Arrange
+    entry = f'''*** Heading
+    :PROPERTIES:
+    :ID:       42-42-42
+    :END:
+    \t\r\nBody Line 1\n
+    '''
+    orgfile = create_file(tmp_path, entry)
+
+    # Act
+    # Extract Entries from specified Org files
+    entries = extract_org_entries(org_files=[orgfile])
+
+    # Process Each Entry from All Notes Files
+    jsonl_string = convert_org_entries_to_jsonl(entries)
+    jsonl_data = [json.loads(json_string) for json_string in jsonl_string.splitlines()] 
+
+    # Assert
+    assert len(jsonl_data) == 1
+
+
+# Helper Functions
+def create_file(tmp_path, entry, filename="test.org"):
+    org_file = tmp_path / f"notes/{filename}"
+    org_file.parent.mkdir()
+    org_file.touch()
+    org_file.write_text(entry)
+    return org_file
\ No newline at end of file