Scratch folder for temporary data moves when using profiles (#112)

CHANGELOG.md

@@ -15,6 +15,15 @@ the top. Non pre-release versions sometimes have an associated name.
 ### Added
 - Added the `--dry-run` option to `kerblam data clean` to only show the files
   that will be deleted without actually doing anything to them.
+- When using a profile, specifying `"_"` as target will simply temporarily hide
+  the file for the duration of the workflow.
+  For example `"test.txt" = "_"` will temporarily remove the `"test.txt"` file
+  from its original position (it will be moved to `.kerblam/scratch`).
+
+### Changed
+- The way that profiles are handled was changed.
+  Now, the original files are moved to `.kerblam/scratch/` during the workflow,
+  instead of remaining in the original directory (and being renamed `.original`).
 
 ## [v1.1.1] - 2024-10-14
 

Cargo.toml

@@ -48,6 +48,7 @@ flate2 = "1.0.28"
 homedir = "0.2.1"
 indicatif = "^0.17"
 log = "^0.4"
+rand = "0.8.5"
 reqwest = { version = "^0.11", default-features = false, features = ["json", "blocking", "rustls-tls"] }
 serde = { version = "^1.0", features = ["derive"] }
 serde_json = "1.0.115"

docs/src/manual/run.md

@@ -89,12 +89,12 @@ kerblam run process_csv --profile alternate
 > by default `data/in`.
 
 Under the hood, Kerblam! will:
-- Rename `input.csv` to `input.csv.original`;
+- Move `input.csv` to a temporary directory in the root of the project named `.kerblam/scratch`,
+  adding a very small salt string to its name (to avoid potential name collisions);
 - Move `different_input.csv` to `input.csv`;
 - Run the analysis as normal;
-- When the run ends (it finishes, it crashes or you kill it), Kerblam! will undo both actions:
-  it moves `different_input.csv` back to its original place and
-  renames `input.csv.original` back to `input.csv`.
+- When the run ends (it finishes, it crashes or you kill it), Kerblam! will restore the original state:
+  it moves both `different_input.csv` and `input.csv.<salt>` back to their original places.
 
 This effectively causes the workflow to run with different input data.
 

src/commands/new.rs

@@ -8,6 +8,7 @@ use crate::VERSION;
 pub fn create_kerblam_project(dir: &PathBuf) -> Result<()> {
     let dirs_to_create: Vec<&str> = vec![
         "",
+        "./.kerblam",
         "./data/in",
         "./data/out",
         "./src/workflows",
@@ -35,6 +36,9 @@ pub fn create_kerblam_project(dir: &PathBuf) -> Result<()> {
     let mut commands_to_run: Vec<(&str, Vec<String>)> = vec![];
     commands_to_run.push(("git", vec![String::from("init")]));
     let mut gitignore_content: Vec<String> = vec![];
+    // We always ignore the .kerblam directory
+    gitignore_content.push(".kerblam".to_string());
+
     // Ask for user input
     // I defined `dirs_to_create` before so that if we ever have to add to them
     // dynamically we can do so here.

src/execution.rs

@@ -441,11 +441,11 @@ impl FileMover {
         }
     }
 
-    pub fn get_from(self) -> PathBuf {
+    pub fn get_from(&self) -> PathBuf {
         self.from.clone()
     }
     #[allow(dead_code)]
-    pub fn get_to(self) -> PathBuf {
+    pub fn get_to(&self) -> PathBuf {
         self.to.clone()
     }
 }