datapod: add hash column to index (#12)

Signed-off-by: Nico Wagner <n.wagner@dnb.de>

crates/datapod/Cargo.toml

@@ -15,6 +15,7 @@ polars.workspace = true
 rayon.workspace = true
 semver.workspace = true
 serde.workspace = true
+sha2 = "0.10.8"
 thiserror.workspace = true
 toml.workspace = true
 

crates/datapod/src/commands/index.rs

@@ -43,6 +43,7 @@ struct Row {
     idn: String,
     path: PathBuf,
     len: u64,
+    hash: String,
 }
 
 impl TryFrom<&PathBuf> for Row {
@@ -54,6 +55,7 @@ impl TryFrom<&PathBuf> for Row {
             idn: doc.idn(),
             path: path.into(),
             len: doc.len(),
+            hash: doc.hash(),
         })
     }
 }
@@ -89,19 +91,22 @@ pub(crate) fn execute(args: Index) -> DatapodResult<()> {
     let mut remote: Vec<&str> = vec![];
     let mut path: Vec<String> = vec![];
     let mut len: Vec<u64> = vec![];
+    let mut hash: Vec<String> = vec![];
 
     for row in rows.into_iter() {
         idn.push(row.idn);
         remote.push(&config.metadata.name);
         path.push(relpath(&row.path, base_dir));
         len.push(row.len);
+        hash.push(row.hash[0..12].to_string());
     }
 
     let mut df = DataFrame::new(vec![
         Series::new("idn", idn),
         Series::new("remote", remote),
         Series::new("path", path),
         Series::new("len", len),
+        Series::new("hash", hash),
     ])?;
 
     match args.output {

crates/datapod/src/document.rs

@@ -1,8 +1,10 @@
+use std::fmt::Write;
 use std::fs::File;
 use std::io::Read;
 use std::path::{Path, PathBuf};
 
 use bstr::BString;
+use sha2::{Digest, Sha256};
 
 use crate::error::DatapodResult;
 
@@ -33,15 +35,20 @@ impl Document {
     }
 
     /// Returns the length of the document in bytes.
-    ///
-    /// ```rust
-    /// use document_stats::Document;
-    ///
-    /// let doc = Document::new("a ∉ ℕ");
-    /// assert_eq!(doc.len(), 9);
-    /// ```
     #[inline]
     pub(crate) fn len(&self) -> u64 {
         self.buf.len() as u64
     }
+
+    /// Returns the SHA256 digest of the document.
+    pub(crate) fn hash(&self) -> String {
+        let mut hasher = Sha256::new();
+        hasher.update(&self.buf);
+
+        let hash = hasher.finalize();
+        hash.iter().fold(String::new(), |mut out, b| {
+            let _ = write!(out, "{b:02x}");
+            out
+        })
+    }
 }