From a4647d345db8a3dd04c63547ad30b5763b52fabe Mon Sep 17 00:00:00 2001
From: Martin Bernstorff <martinbernstorff@gmail.com>
Date: Sun, 27 Oct 2024 14:19:41 +0100
Subject: [PATCH] feat: improve logging (#746)

---
 memium/__main__.py               |  6 +++++-
 memium/source/document_source.py | 18 ++++++++++++++----
 2 files changed, 19 insertions(+), 5 deletions(-)

diff --git a/memium/__main__.py b/memium/__main__.py
index de47776..c694280 100644
--- a/memium/__main__.py
+++ b/memium/__main__.py
@@ -1,6 +1,7 @@
 import logging
 import sys
 import time
+from datetime import datetime
 from functools import partial
 from pathlib import Path
 from typing import Annotated, Optional
@@ -53,6 +54,7 @@ def cli(
         bool, typer.Option(help="Skip all syncing, useful for smoketesting of the interface")
     ] = False,
 ):
+    start_time = datetime.now()
     config_dir = input_dir / ".memium"
     config_dir.mkdir(exist_ok=True)
 
@@ -81,7 +83,9 @@ def cli(
     main_fn()
 
     if watch_seconds:
-        log.info(f"Sync complete, sleeping for {watch_seconds} seconds")
+        log.info(
+            f"Sync complete in {(datetime.now() - start_time).total_seconds()} seconds, sleeping for {watch_seconds} seconds"
+        )
         time.sleep(watch_seconds)
         main_fn()
 
diff --git a/memium/source/document_source.py b/memium/source/document_source.py
index 3e1eecf..c1c90b2 100644
--- a/memium/source/document_source.py
+++ b/memium/source/document_source.py
@@ -17,6 +17,9 @@ class FileNotRetrievedError(Exception):
     path: Path
     error: Exception
 
+    def __repr__(self) -> str:
+        return f"{self.path}: Could not retrieve due to '{self.error}'"
+
 
 class BaseDocumentSource(Protocol):
     def get_documents(self) -> Sequence[Document]: ...
@@ -59,10 +62,17 @@ def _sanitize_to_valid_markdown(self, input_str: str) -> str:
 
     def _get_document_from_file(self, file_path: Path) -> Document | FileNotRetrievedError:
         try:
-            with file_path.open("r", encoding="utf8") as f:
-                return Document(
-                    content=self._sanitize_to_valid_markdown(f.read()), source_path=file_path
-                )
+            try:
+                contents = file_path.read_text(encoding="utf8")
+            except Exception as e:
+                raise Exception(f"Could not read file {file_path}") from e
+
+            try:
+                sanitized = self._sanitize_to_valid_markdown(contents)
+            except Exception as e:
+                raise Exception(f"Could not sanitize file {file_path}") from e
+
+            return Document(content=sanitized, source_path=file_path)
         except Exception as e:
             log.warning(f"Could not retrieve {file_path}: {e}")
             return FileNotRetrievedError(file_path, e)