From c8c142bd1b06588126aa957ad76b42327165f4e7 Mon Sep 17 00:00:00 2001
From: Jon Michael Aanes <jonjmaa@gmail.com>
Date: Wed, 16 Oct 2024 23:58:22 +0200
Subject: [PATCH] Obsidian import: More intelligent time behaviour

---
 obsidian_import/__init__.py  |  6 +++---
 obsidian_import/obsidian.py  | 28 +++++++++++++++++++++++-----
 personal_data/activity.py    | 11 +++++++++++
 test/test_obsidian_format.py | 18 ++++++++++++++++++
 4 files changed, 55 insertions(+), 8 deletions(-)
 create mode 100644 test/test_obsidian_format.py

diff --git a/obsidian_import/__init__.py b/obsidian_import/__init__.py
index e8cb101..ee775a4 100644
--- a/obsidian_import/__init__.py
+++ b/obsidian_import/__init__.py
@@ -119,8 +119,9 @@ def import_watched_series_csv(vault: ObsidianVault, rows: Rows) -> int:
             sample.single_label_with_category('episode.index'),
             sample.single_label_with_category('episode.name'),
         )
-        return Event(sample.start_at.time(),
-                     sample.end_at.time(),
+        expected_tz = datetime.timezone(datetime.timedelta(hours=2)) # TODO: Determine this in a more intelligent manner
+        return Event(sample.start_at.astimezone(expected_tz).replace(second=0,microsecond=0).time(),
+                     sample.end_at.astimezone(expected_tz).replace(second=0,microsecond=0).time(),
                      verb,
                      sample.single_label_with_category('series.name'),
                      comment,
@@ -162,6 +163,5 @@ def import_data(obsidian_path: Path, dry_run=True):
         data_path = Path('output/show_episodes_watched.csv')
         rows = load_csv_file(data_path)
         logger.info('Loaded CSV with %d lines', len(rows))
-        rows = rows[:7]
         num_updated = import_watched_series_csv(vault, rows)
         logger.info('Updated %d files', num_updated)
diff --git a/obsidian_import/obsidian.py b/obsidian_import/obsidian.py
index b0dc767..be2085b 100644
--- a/obsidian_import/obsidian.py
+++ b/obsidian_import/obsidian.py
@@ -210,25 +210,43 @@ def format_event_string(event: Event) -> str:
     ):
         return event.comment
 
-    return f'{event.start_time:%H:%M} | {event.verb} [[{event.subject}]]. {event.comment}'.strip()
+    buf = []
+    buf.append(f'{event.start_time:%H:%M}')
+    if event.end_time and event.end_time != event.start_time:
+        buf.append(f'-{event.end_time:%H:%M}')
+    buf.append(' | ')
+    buf.append(event.verb)
+    buf.append(' [[')
+    buf.append(event.subject)
+    buf.append(']]. ')
+    buf.append(event.comment.strip())
+
+    return ''.join(buf)
 
 
 RE_TIME = r'(\d\d:\d\d(?::\d\d(?:\.\d+?))?)'
+RE_VERB = r'(\w+(?:ed|te))'
+RE_LINK_MD = r'\[([^\]]*)\]\(?:[^)]*\)'
+RE_LINK_WIKI = r'\[\[([^\]]*)\]\]'
+
+RE_TIME_FORMAT = RE_TIME + r'(?:\s*\-\s*' + RE_TIME + r')?'
 
 
 def parse_event_string(event_str: str) -> Event:
     if m := re.match(
-        r'^\s*' + RE_TIME + r'[ :\|-]*(\w+ed)\s+\[([^\]]*)\]\([^)]*\)\.?\s*(.*)$',
+        r'^\s*' + RE_TIME_FORMAT + r'[ :\|-]*'+RE_VERB+r'\s+'+RE_LINK_MD+r'\.?\s*(.*)$',
         event_str,
     ):
         start = datetime.time.fromisoformat(m.group(1))
-        return Event(start, start, m.group(2), m.group(3), m.group(4))
+        end = datetime.time.fromisoformat(m.group(2)) if m.group(2) else start
+        return Event(start, end, m.group(3), m.group(4), m.group(5))
     if m := re.match(
-        r'^\s*' + RE_TIME + '[ :\|-]*(\w+ed)\s+\[\[([^\]]*)\]\]\.?\s*(.*)$',
+            r'^\s*' + RE_TIME_FORMAT + r'[ :\|-]*'+RE_VERB+r'\s+'+RE_LINK_WIKI+r'\.?\s*(.*)$',
         event_str,
     ):
         start = datetime.time.fromisoformat(m.group(1))
-        return Event(start, start, m.group(2), m.group(3), m.group(4))
+        end = datetime.time.fromisoformat(m.group(2)) if m.group(2) else start
+        return Event(start, end, m.group(3), m.group(4), m.group(5))
     logger.info('Could not parse format: %s', event_str)
     return Event(None, None, None, None, event_str)
 
diff --git a/personal_data/activity.py b/personal_data/activity.py
index 89e70cd..ba13826 100644
--- a/personal_data/activity.py
+++ b/personal_data/activity.py
@@ -23,6 +23,10 @@ class ActivitySample:
     start_at: datetime.datetime | None
     end_at: datetime.datetime | None
 
+    def __post_init__(self):
+        if self.start_at and self.end_at:
+            assert self.start_at <= self.end_at
+
     def single_label_with_category(self, category: str) -> str:
         for label in self.labels:
             if label.category == category:
@@ -35,6 +39,11 @@ class RealizedActivitySample(ActivitySample):
     start_at: datetime.datetime
     end_at: datetime.datetime
 
+    def __post_init__(self):
+        assert self.start_at is not None
+        assert self.end_at is not None
+        assert self.start_at <= self.end_at
+
 
 def heuristically_realize_samples(
     samples: list[ActivitySample],
@@ -45,6 +54,8 @@ def heuristically_realize_samples(
     * No samples overlap.
     """
 
+    samples.sort(key = lambda x: x.end_at)
+
     previous_sample_end = None
     for sample in samples:
         end_at = sample.end_at
diff --git a/test/test_obsidian_format.py b/test/test_obsidian_format.py
new file mode 100644
index 0000000..86909bf
--- /dev/null
+++ b/test/test_obsidian_format.py
@@ -0,0 +1,18 @@
+import datetime
+import pytest
+
+from obsidian_import import obsidian
+
+
+EXAMPLES = [
+        obsidian.Event(datetime.time(12, 0, 0), datetime.time(12, 0, 0), "Ate",
+                       "Lunch", "instantly"),
+        obsidian.Event(datetime.time(20, 0, 0), datetime.time(22, 0, 0),
+                       "Watched", "Tom and Jerry", "on the *Television*"),
+        obsidian.Event(None, None, None, None, "Took a walk"),
+]
+
+@pytest.mark.parametrize("event", EXAMPLES)
+def test_format_preserves_information(event: obsidian.Event):
+    formatted = obsidian.format_event_string(event)
+    assert obsidian.parse_event_string(formatted) == event