update webtable post

Author: marvin-j97

src/content/blog/2025-03-14_webtable.md

@@ -7,11 +7,12 @@ tags:
   - wide column
   - data modelling
 published_at: 2025-03-21T15:55:27.431Z
-last_modified_at: 2025-03-21T15:55:27.431Z
+last_modified_at: 2025-08-04T23:55:27.431Z
 image: /media/thumbs/google.jpg
 ---
 
-Everyone knows Google. For almost the entire’s internet lifetime, the Google search page has been the starting point of almost any internet user’s surfing session.
+Everyone knows Google.
+For almost the entire’s internet lifetime, the Google search page has been the starting point of almost any internet user’s surfing session.
 To be able to serve search requests of (almost) anything that can be found on the internet, Google has to manage a huge search index.
 Storing a large portion of the internet requires petabytes, [if not exabytes](https://cloud.google.com/blog/topics/developers-practitioners/how-big-cloud-bigtable?hl=en), of storage.
 
@@ -39,7 +40,7 @@ The row key defines how the table is ordered, so it is the only indexing mechani
 
 ## The descendants of Bigtable: LevelDB & RocksDB
 
-LevelDB, an embeddable LSM-tree based key-value storage engine by Google, was released in 2011 and a [direct descendent of Bigtable code](https://opensource.googleblog.com/2011/07/leveldb-fast-persistent-key-value-store.html).
+LevelDB, an embeddable LSM-tree based key-value storage engine by Google, was released in 2011 and is a [direct descendent of Bigtable code](https://opensource.googleblog.com/2011/07/leveldb-fast-persistent-key-value-store.html).
 It is now in [maintenance-only mode](https://github.com/google/leveldb/commit/aa5479bbf47e9df86e0afbb89e6246085f22cdd4).
 
 > Fun fact: [Every Chromium browser ships with LevelDB](https://chromium.googlesource.com/chromium/src/+/refs/heads/main/content/browser/indexed_db/indexed_db_leveldb_coding.cc).
@@ -108,9 +109,9 @@ And to retrieve a row, we do:
 for kv in webtable.prefix("jupiter\0") {
     let (k, v) = kv?;
     let mut splits = k.split(|&x| x == b'\0');
-    let row_key = std::str::from_utf8(splits.next().unwrap()).unwrap();
-    let col_fam = std::str::from_utf8(splits.next().unwrap()).unwrap();
-    let col_qua = std::str::from_utf8(splits.next().unwrap()).unwrap();
+    let row_key = std::str::from_utf8(splits.next()?)?;
+    let col_fam = std::str::from_utf8(splits.next()?)?;
+    let col_qua = std::str::from_utf8(splits.next()?)?;
 
     let mut buf = [0; std::mem::size_of::<Timestamp>()];
     buf.copy_from_slice(&k[k.len() - std::mem::size_of::<Timestamp>()..]);
@@ -200,10 +201,10 @@ impl WideColumnTable {
             // ignore the TableCell::new syntax, it's some nasty self_cell stuff
             Ok(TableCell::new(kv?, |(k, v)| {
                 let mut splits = k.split(|&x| x == b'\0');
-                let row_key = std::str::from_utf8(splits.next().unwrap()).unwrap();
-                let column_family = std::str::from_utf8(splits.next().unwrap()).unwrap();
-                let column_qualifier = std::str::from_utf8(splits.next().unwrap()).unwrap();
-                let ts_bytes = splits.next().unwrap();
+                let row_key = std::str::from_utf8(splits.next()?)?;
+                let column_family = std::str::from_utf8(splits.next()?)?;
+                let column_qualifier = std::str::from_utf8(splits.next()?)?;
+                let ts_bytes = splits.next()?;
 
                 let mut buf = [0; std::mem::size_of::<Timestamp>()];
                 buf.copy_from_slice(ts_bytes);
@@ -449,10 +450,10 @@ let test_urls = ["https://vedur.is", "https://news.ycombinator.com"];
 
 for url in test_urls {
     eprintln!("Scraping {url:?}");
-    let res = reqwest::blocking::get(url).unwrap();
+    let res = reqwest::blocking::get(url)?;
 
     if res.status().is_success() {
-        let html = res.text().unwrap();
+        let html = res.text()?;
         webtable.insert(url, &html)?;
     }
 }