feat: add list of available profiles to run error

CHANGELOG.md

@@ -12,7 +12,10 @@ Versions are listed in reverse chronological order, with the most recent at
 the top. Non pre-release versions sometimes have an associated name.
 
 ## [Unreleased]
-- Nothing yet!
+### New
+- The error message you get when running `kerblam run` with no parameters now
+  includes a list of available profiles, or tells you that you have specified
+  no profiles.
 
 ## [v1.0.0-rc.3] - 2024-06-24
 

Cargo.toml

@@ -77,4 +77,5 @@ chwd = "0.2.0"
 git2 = "0.18.2"
 paste = "1.0.14"
 rusty-fork = "0.3.0"
+serial_test = "3.1.1"
 similar = "2.4.0"

docs/src/manual/pipe_docstrings.md

@@ -2,10 +2,12 @@
 If you execute `kerblam run` without specifying a pipe (or you try to run a 
 pipe that does not exist), you will get a message like this:
 ```
-Error: no runtime specified. Available runtimes:
+Error: No runtime specified. Available runtimes:
     ◾◾ process_csv
     🐋◾ save_plots
     ◾◾ generate_metrics
+
+Available profiles: No profiles defined.
 ```
 The whale emoji (🐋) represents pipes that [have an associated Docker container](run_containers.html).
 
@@ -32,10 +34,12 @@ to only have a single contiguous description block in each file.
 
 The output of `kerblam run` will now read:
 ```
-Error: no runtime specified. Available runtimes:
+Error: No runtime specified. Available runtimes:
     ◾📜 process_csv :: Calculate the sums of the input metrics
     🐋◾ save_plots
     ◾◾ generate_metrics
+
+Available profiles: No profiles defined.
 ```
 The scroll (📜) emoji appears when Kerblam! notices a long description.
 You can show the full description for such pipes with `kerblam run process_csv --desc`.

docs/src/manual/run.md

@@ -31,6 +31,24 @@ In short, `kerblam run` does something similar to this:
 This is why workflows are written as if they are executed in the root of the
 project, because they are.
 
+### Listing out workflows
+If you just want a list of workflows that Kerblam! can see, just use
+`kerblam run` with no workflow specified. Kerblam will reply with something
+like this:
+
+```
+Error: No runtime specified. Available runtimes:
+    ◾📜 process_csv :: Calculate the sums of the input metrics
+    🐋◾ save_plots
+    ◾◾ generate_metrics
+
+Available profiles: No profiles defined.
+```
+
+Workflows with a 📜 have [an associated description](pipe_docstrings.md), and 
+those with a 🐋 have [an associated docker container](run_containers.md).
+You also get a list of available data profiles, which are detailed just below.
+
 ## Data Profiles - Running the same workflows on different data
 
 You can run your same workflows, *as-is*, on different data thanks to data profiles.
@@ -160,4 +178,3 @@ For example, you can tell `make` to build a different target with this syntax:
 kerblam run make_workflow -- other_target
 ```
 As if you had run `make other_target` yourself.
-

docs/src/manual/run_containers.md

@@ -55,6 +55,8 @@ are prepended with a little whale (🐋):
 Error: No runtime specified. Available runtimes:
     🐋◾ my_workflow :: Generate the output data in a docker container
     ◾◾ local_workflow :: Run some code locally
+
+Available profiles: No profiles defined.
 ```
 
 ### Default dockerfile
@@ -68,6 +70,8 @@ pipes that use the default container, so you can identify them easily:
 Error: No runtime specified. Available runtimes:
     🐋◾ my_workflow :: Generate the output data in a docker container
     🐟◾ another :: Run in the default container
+
+Available profiles: No profiles defined.
 ```
 
 ### Switching backends

src/commands/run.rs

@@ -1,5 +1,6 @@
+use crate::options::extract_profile_paths;
 use std::collections::HashMap;
-use std::path::{Path, PathBuf};
+use std::path::PathBuf;
 
 use crate::cache::{check_last_profile, delete_last_profile, get_cache};
 use crate::execution::{setup_ctrlc_hook, Executor, FileMover};
@@ -9,148 +10,6 @@ use crate::utils::update_timestamps;
 
 use anyhow::{anyhow, bail, Result};
 
-/// Push a bit of a string to the end of this path
-///
-/// Useful if you want to add an extension to the path.
-/// Requires a clone.
-fn push_fragment(buffer: impl AsRef<Path>, ext: &str) -> PathBuf {
-    let buffer = buffer.as_ref();
-    let mut path = buffer.as_os_str().to_owned();
-    path.push(ext);
-    path.into()
-}
-
-fn infer_test_data(paths: Vec<PathBuf>) -> HashMap<PathBuf, PathBuf> {
-    let mut matches: HashMap<PathBuf, PathBuf> = HashMap::new();
-
-    for path in paths.clone() {
-        let file_name = path.file_name().unwrap().to_string_lossy();
-        if file_name.starts_with("test_") {
-            let slug = file_name.trim_start_matches("test_");
-            let potential_target = path.clone().with_file_name(slug);
-            if paths.iter().any(|x| *x == potential_target) {
-                matches.insert(potential_target, path);
-            }
-        }
-    }
-
-    matches
-}
-
-// TODO: This checks for the existence of profile paths here. This is a bad
-// thing. It's best to handle the error when we actually do the move.
-// This was done this way because I want a nice error list.
-// The 'check_existence' check was added to overcome this, but it's a hack.
-fn extract_profile_paths(
-    config: &KerblamTomlOptions,
-    profile_name: &str,
-    check_existance: bool,
-) -> Result<Vec<FileMover>> {
-    let root_dir = config.input_data_dir();
-
-    // If there are no profiles, an empty hashmap is OK intead:
-    // we can add the default "test" profile anyway.
-    let mut profiles = {
-        let data = config.clone().data;
-        match data {
-            Some(x) => x.profiles.unwrap_or(HashMap::new()),
-            None => HashMap::new(),
-        }
-    };
-
-    // add the default 'test' profile
-    if !profiles.keys().any(|x| x == "test") {
-        let input_files = config.input_files();
-        let inferred_test = infer_test_data(input_files);
-        if !inferred_test.is_empty() {
-            log::debug!("Inserted inferred test profile: {inferred_test:?}");
-            profiles.insert("test".to_string(), inferred_test);
-        }
-    }
-
-    let profile = profiles
-        .get(profile_name)
-        .ok_or(anyhow!("Could not find {} profile", profile_name))?;
-
-    // Check if the sources exist, otherwise we crash now, and not later
-    // when we actually move the files.
-    let exist_check: Vec<anyhow::Error> = profile
-        .iter()
-        .flat_map(|(a, b)| [a, b])
-        .map(|file| {
-            let f = &root_dir.join(file);
-            log::debug!("Checking if {f:?} exists...");
-            match f.try_exists() {
-                Ok(i) => {
-                    if i {
-                        Ok(())
-                    } else {
-                        bail!("\t - {file:?} does not exist!")
-                    }
-                }
-                Err(e) => bail!("\t- {file:?} - {e:?}"),
-            }
-        })
-        .filter_map(|x| x.err())
-        .collect();
-
-    if !exist_check.is_empty() & check_existance {
-        let mut missing: Vec<String> = Vec::with_capacity(exist_check.len());
-        for item in exist_check {
-            missing.push(item.to_string());
-        }
-        bail!(
-            "Failed to find some profiles files:\n{}",
-            missing.join("\n")
-        )
-    }
-
-    // Also check if the targets do NOT exist, so we don't overwrite anything
-    let exist_check: Vec<anyhow::Error> = profile
-        .iter()
-        .flat_map(|(a, b)| [a, b])
-        .map(|file| {
-            let f = &root_dir.join(push_fragment(file, ".original"));
-            log::debug!("Checking if {f:?} destroys files...");
-            if f.exists() {
-                bail!("\t- {:?} would be destroyed by {:?}!", f, file)
-            };
-            Ok(())
-        })
-        .filter_map(|x| x.err())
-        .collect();
-
-    if !exist_check.is_empty() & check_existance {
-        let mut missing: Vec<String> = Vec::with_capacity(exist_check.len());
-        for item in exist_check {
-            missing.push(item.to_string());
-        }
-        bail!(
-            "Some profile temporary files would overwrite real files:\n{}",
-            missing.join("\n")
-        )
-    }
-
-    Ok(profile
-        .iter()
-        .flat_map(|(original, profile)| {
-            // We need two FileMovers. One for the temporary file
-            // that holds the original file (e.g. 'to'), and one for the
-            // profile-to-original rename.
-            // To unwind, we just redo the transaction, but in reverse.
-            [
-                // This one moves the original to the temporary file
-                FileMover::from((
-                    &root_dir.join(original),
-                    &root_dir.join(push_fragment(original, ".original")),
-                )),
-                // This one moves the profile one to the original one
-                FileMover::from((&root_dir.join(profile), &root_dir.join(original))),
-            ]
-        })
-        .collect())
-}
-
 pub fn kerblam_run_project(
     config: KerblamTomlOptions,
     pipe: Pipe,