feat: add collapsible highlight with context to output (#24)

- [ ] I have considered whether this PR needs review, and requested a review if necessary.

Fixes issue #

# Notes for reviewers
Reviewers can skip X, but should pay attention to Y.

memorymarker/__main__.py

@@ -26,16 +26,26 @@ class TimestampHandler:
     filepath: Path
 
     def update_timestamp(self) -> None:
+        if not self.filepath.exists():
+            self.filepath.touch()
+
         self.filepath.write_text(dt.datetime.now(pytz.UTC).isoformat())
 
-    def get_timestamp(self) -> dt.datetime:
-        return dt.datetime.fromisoformat(self.filepath.read_text())
+    def get_timestamp(self) -> dt.datetime | None:
+        try:
+            dt.datetime.fromisoformat(self.filepath.read_text())
+        except FileNotFoundError:
+            return None
 
 
 @app.command()
 def typer_cli(
     output_dir: Path = typer.Argument(
-        Path("questions"), help="Directory to save the generated questions to"
+        Path("questions"),
+        help="Directory to save the generated questions to",
+        file_okay=False,
+        dir_okay=True,
+        writable=True,
     ),
     max_n: int = typer.Argument(
         1, help="Maximum number of questions to generate from highlights"
@@ -44,17 +54,22 @@ def typer_cli(
         True, help="Only generate questions from highlights since last run"
     ),
 ) -> None:
+    output_dir.mkdir(exist_ok=True, parents=True)
     last_run_timestamper = TimestampHandler(output_dir / ".memorymarker")
-    typer.echo(
-        f"Last run at UTC {last_run_timestamper.get_timestamp().strftime('%Y-%m-%d %H:%M:%S')}"
-    )
+    last_run_timestamp = last_run_timestamper.get_timestamp()
+
     typer.echo("Fetching new highlights...")
     highlights = Omnivore().get_highlights()
 
     if only_new:
-        highlights = highlights.filter(
-            lambda _: _.updated_at > last_run_timestamper.get_timestamp()
+        if not last_run_timestamp:
+            typer.echo("No last run timestamp found, exiting")
+            return
+
+        typer.echo(
+            f"Last run at UTC {last_run_timestamp.strftime('%Y-%m-%d %H:%M:%S')}"
         )
+        highlights = highlights.filter(lambda _: _.updated_at > last_run_timestamp)
         last_run_timestamper.update_timestamp()
 
         if highlights.count() == 0:

memorymarker/highlight_providers/base.py

@@ -16,11 +16,23 @@ class OrphanHighlight:
 class HydratedHighlight(BaseModel):
     source_doc_title: str
     source_doc_uri: str
+
+    prefix: str | None
     highlighted_text: str
-    context: str
+    suffix: str | None
+
     source_highlight_uri: str | None = None
     updated_at: dt.datetime
 
+    @property
+    def context(self) -> str:
+        context = ""
+        context += self.prefix or ""
+        context += self.highlighted_text
+        context += self.suffix or ""
+
+        return context
+
 
 class HighlightSource(Protocol):
     def get_highlights(self) -> tuple[OrphanHighlight]:

memorymarker/highlight_providers/hydrator/main.py

@@ -100,7 +100,8 @@ def hydrate_highlights(
                     highlighted_text=highlight.highlight,
                     source_doc_uri=highlight.uri,
                     source_doc_title=highlight.title,
-                    context=context,
+                    prefix=context[:100],
+                    suffix=context[-100:],
                 ),  # type: ignore
             )
 

memorymarker/highlight_providers/omnivore.py

@@ -20,7 +20,8 @@ def _parse_highlight(self, highlight: Mapping[str, str]) -> HydratedHighlight:
             source_doc_title=self.title,
             source_doc_uri=self.uri,
             highlighted_text=highlight["quote"],
-            context=f"{highlight['prefix']} {highlight['quote']} {highlight['suffix']}",
+            prefix=highlight["prefix"],
+            suffix=highlight["suffix"],
             updated_at=highlight["updatedAt"],  # type: ignore
             source_highlight_uri=f"https://omnivore.app/me/{self.slug}#{highlight["id"]}",
         )

memorymarker/persist_questions/__snapshots__/test_markdown.ambr

@@ -3,7 +3,10 @@
   '''
   Q. What is the meaning of life?
   A. 42
-  [Highlight](https://en.wikipedia.org/wiki/The_Hitchhiker%27s_Guide_to_the_Galaxy#meaning_of_life)
+
+  > [!NOTE]- Highlight
+  > ==42==
+  > [Link](https://en.wikipedia.org/wiki/The_Hitchhiker%27s_Guide_to_the_Galaxy#meaning_of_life)
 
 
   '''

memorymarker/persist_questions/markdown.py

@@ -1,7 +1,6 @@
 import re
 from pathlib import Path
 
-
 from memorymarker.question_generator.question_generator import QAPrompt
 
 
@@ -11,15 +10,18 @@ def clean_filename(filename: str) -> str:
 
 
 def q_to_markdown(prompt: QAPrompt) -> str:
+    highlight = prompt.hydrated_highlight
     return f"""Q. {prompt.question}
 A. {prompt.answer}
-[Highlight]({prompt.hydrated_highlight.source_highlight_uri})
+
+> [!NOTE]- Highlight
+> {highlight.prefix or ""}=={highlight.highlighted_text}=={highlight.suffix.strip() if highlight.suffix is not None else ""}
+> [Link]({highlight.source_highlight_uri})
 \n"""
 
 
 def write_md(contents: str, file_title: str, save_dir: Path) -> None:
     """Write markdown to file. Append if exists"""
-    save_dir.mkdir(exist_ok=True, parents=True)
     with (save_dir / f"{clean_filename(file_title)}.md").open(mode="a") as f:
         f.write(contents + "\n")
 

memorymarker/persist_questions/test_markdown.py

@@ -17,6 +17,8 @@ class FakeHydratedHighlight(HydratedHighlight):
     context: str = "The meaning of life is 42"
     source_highlight_uri: str = "https://en.wikipedia.org/wiki/The_Hitchhiker%27s_Guide_to_the_Galaxy#meaning_of_life"
     updated_at: dt.datetime = dt.datetime.now()
+    prefix: str = ""
+    suffix: str = ""
 
 
 @pytest.fixture(scope="module")

memorymarker/question_generator/test_highlights_to_questions.py

@@ -15,7 +15,8 @@ def model() -> h2q.ChatOpenAI:
 @pytest.fixture(scope="module")
 def hydrated_highlight() -> HydratedHighlight:
     return HydratedHighlight(
-        context="Mitochondria is the powerhouse of the cell",
+        prefix="",
+        suffix=" is the powerhouse of the cell",
         highlighted_text="Mitochondria",
         source_doc_uri="https://en.wikipedia.org/wiki/Mitochondrion",
         source_doc_title="Mitochondrion - Wikipedia",
@@ -40,14 +41,16 @@ async def test_model_response(
 async def test_multi_response(model: h2q.ChatOpenAI) -> None:
     highlights = [
         HydratedHighlight(
-            context="Mitochondria is the powerhouse of the cell",
+            prefix="",
+            suffix=" is the powerhouse of the cell",
             highlighted_text="Mitochondria",
             source_doc_uri="https://en.wikipedia.org/wiki/Mitochondrion",
             source_doc_title="Mitochondrion - Wikipedia",
             updated_at=datetime.now(),
         ),
         HydratedHighlight(
-            context="The first rule of Fight Club is that you don't talk about Fight Club",
+            prefix="The first rule of ",
+            suffix=" is that you don't talk about Fight Club",
             highlighted_text="Fight Club",
             source_doc_uri="https://en.wikipedia.org/wiki/Fight_Club",
             source_doc_title="Fight Club - Wikipedia",