From 79a89125edd6529177b00bca8d09fc9e9fb898a5 Mon Sep 17 00:00:00 2001
From: Lucas <zhangyijunmetro@hotmail.com>
Date: Thu, 8 Feb 2024 11:08:47 +0800
Subject: [PATCH] Add `mbtiles diff` command (#1068)

Implement `mbtiles diff a.mbtiles b.mbtiles diff.mbtiles` command

This should behave exactly the same as `mbtiles copy a.mbtiles
--diff-with-file b.mbtiles diff.mbtiles`.

---------

Co-authored-by: Yuri Astrakhan <yuriastrakhan@gmail.com>
---
 docs/src/mbtiles-diff.md   | 26 +++++++++++++++++++
 mbtiles/src/bin/mbtiles.rs | 52 +++++++++++++++++++++++++++++++++++++-
 tests/test.sh              |  6 ++++-
 3 files changed, 82 insertions(+), 2 deletions(-)
 create mode 100644 docs/src/mbtiles-diff.md

diff --git a/docs/src/mbtiles-diff.md b/docs/src/mbtiles-diff.md
new file mode 100644
index 00000000..9f24356c
--- /dev/null
+++ b/docs/src/mbtiles-diff.md
@@ -0,0 +1,26 @@
+# Diffing MBTiles
+
+## `mbtiles diff`
+
+Diff command compares two mbtiles files `A` and `B`, and generates a diff (delta) file.
+If the diff file is [applied](mbtiles-copy.md#mbtiles-apply-patch) to `A`, it will produce `B`.  
+The diff file will contain all tiles that are different between the two files
+(modifications, insertions, and deletions as `NULL` values), for both the tile and metadata tables.  
+The only exception is `agg_tiles_has` metadata value. It will be renamed to `agg_tiles_hash_in_diff` and a
+new `agg_tiles_hash` will be generated for the diff file itself.
+
+```shell
+# This command will comapre `a.mbtiles` and `b.mbtiles`, and generate a new diff file `diff.mbtiles`.
+mbtiles diff a.mbtiles b.mbtiles diff.mbtiles
+
+# If diff.mbtiles is applied to a.mbtiles, it will produce b.mbtiles 
+mbtiles apply-diff a.mbtiles diff.mbtiles b2.mbtiles
+
+# b.mbtiles and b2.mbtiles should now be the same
+# Validate both files and see that their hash values are identical
+mbtiles validate b.mbtiles
+[INFO ] The agg_tiles_hashes=E95C1081447FB25674DCC1EB97F60C26 has been verified for b.mbtiles
+
+mbtiles validate b2.mbtiles
+[INFO ] The agg_tiles_hashes=E95C1081447FB25674DCC1EB97F60C26 has been verified for b2.mbtiles
+```
diff --git a/mbtiles/src/bin/mbtiles.rs b/mbtiles/src/bin/mbtiles.rs
index 0f907407..bc870239 100644
--- a/mbtiles/src/bin/mbtiles.rs
+++ b/mbtiles/src/bin/mbtiles.rs
@@ -52,6 +52,13 @@ enum Commands {
         /// Value to set, or nothing if the key should be deleted.
         value: Option<String>,
     },
+    /// Compare two files A and B, and generate a new diff file. If the diff file is applied to A, it will produce B.
+    #[command(name = "diff")]
+    Diff {
+        file_a: PathBuf,
+        file_b: PathBuf,
+        diff: PathBuf,
+    },
     /// Copy tiles from one mbtiles file to another.
     #[command(name = "copy", alias = "cp")]
     Copy(CopyArgs),
@@ -212,6 +219,28 @@ async fn main_int() -> anyhow::Result<()> {
             println!("MBTiles file summary for {mbt}");
             println!("{}", mbt.summary(&mut conn).await?);
         }
+        Commands::Diff {
+            file_a,
+            file_b,
+            diff,
+        } => {
+            let opts = MbtilesCopier {
+                src_file: file_a,
+                diff_with_file: Some(file_b),
+                dst_file: diff,
+                copy: CopyType::All,
+                skip_agg_tiles_hash: false,
+                on_duplicate: Some(CopyDuplicateMode::Override),
+                dst_type_cli: None,
+                dst_type: None,
+                min_zoom: None,
+                max_zoom: None,
+                zoom_levels: vec![],
+                bbox: vec![],
+                apply_patch: None,
+            };
+            opts.run().await?;
+        }
     }
 
     Ok(())
@@ -253,7 +282,7 @@ mod tests {
     use mbtiles::CopyDuplicateMode;
 
     use super::*;
-    use crate::Commands::{ApplyPatch, Copy, MetaGetValue, MetaSetValue, Validate};
+    use crate::Commands::{ApplyPatch, Copy, Diff, MetaGetValue, MetaSetValue, Validate};
     use crate::{Args, IntegrityCheckType};
 
     #[test]
@@ -524,4 +553,25 @@ mod tests {
             }
         );
     }
+
+    #[test]
+    fn test_diff() {
+        assert_eq!(
+            Args::parse_from([
+                "mbtiles",
+                "diff",
+                "file-a.mbtiles",
+                "file-b.mbtiles",
+                "../delta.mbtiles",
+            ]),
+            Args {
+                verbose: false,
+                command: Diff {
+                    file_a: PathBuf::from("file-a.mbtiles"),
+                    file_b: PathBuf::from("file-b.mbtiles"),
+                    diff: PathBuf::from("../delta.mbtiles"),
+                }
+            }
+        );
+    }
 }
diff --git a/tests/test.sh b/tests/test.sh
index e53580ce..6371cc76 100755
--- a/tests/test.sh
+++ b/tests/test.sh
@@ -458,7 +458,11 @@ if [[ "$MBTILES_BIN" != "-" ]]; then
     "$TEST_TEMP_DIR/world_cities_diff.mbtiles" \
     --diff-with-file ./tests/fixtures/mbtiles/world_cities_modified.mbtiles \
     2>&1 | tee "$TEST_OUT_DIR/copy_diff.txt"
-
+  $MBTILES_BIN diff \
+       ./tests/fixtures/mbtiles/world_cities.mbtiles \
+       ./tests/fixtures/mbtiles/world_cities_modified.mbtiles \
+       "$TEST_TEMP_DIR/world_cities_diff2.mbtiles" \
+       2>&1 | tee "$TEST_OUT_DIR/copy_diff2.txt"
   if command -v sqlite3 > /dev/null; then
     # Apply this diff to the original version of the file
     cp ./tests/fixtures/mbtiles/world_cities.mbtiles "$TEST_TEMP_DIR/world_cities_copy.mbtiles"