From 628ef55e3bf50c0366baf873347520b82d91a7c2 Mon Sep 17 00:00:00 2001
From: Arpad Borsos <arpad.borsos@sentry.io>
Date: Fri, 20 Dec 2024 13:04:09 +0100
Subject: [PATCH] WIP: experimentally add a filetree implementation

---
 src/filetree/components.rs |  49 ++++++++++++++
 src/filetree/mod.rs        | 130 +++++++++++++++++++++++++++++++++++++
 src/lib.rs                 |   1 +
 3 files changed, 180 insertions(+)
 create mode 100644 src/filetree/components.rs
 create mode 100644 src/filetree/mod.rs

diff --git a/src/filetree/components.rs b/src/filetree/components.rs
new file mode 100644
index 0000000..7ffa609
--- /dev/null
+++ b/src/filetree/components.rs
@@ -0,0 +1,49 @@
+/// Yields path components in reverse order.
+///
+/// This also takes care to normalize `.` and `..` components,
+/// and it skips any trailing `..`.
+pub fn reverse_components(path: &str) -> impl Iterator<Item = &str> {
+    let mut skip = 0;
+    let mut components = path.split('/').rev();
+
+    std::iter::from_fn(move || {
+        while let Some(next) = components.next() {
+            match next {
+                "." => continue,
+                ".." => {
+                    skip += 1;
+                    continue;
+                }
+                _ if skip > 0 => {
+                    skip -= 1;
+                    continue;
+                }
+                component => return Some(component),
+            }
+        }
+        None
+    })
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_reverse_components() {
+        let components: Vec<_> = reverse_components("mod.rs").collect();
+        assert_eq!(components, &["mod.rs"]);
+
+        let components: Vec<_> = reverse_components("./foo/./bar/mod.rs").collect();
+        assert_eq!(components, &["mod.rs", "bar", "foo"]);
+
+        let components: Vec<_> = reverse_components("./foo/../bar/mod.rs").collect();
+        assert_eq!(components, &["mod.rs", "bar"]);
+
+        let components: Vec<_> = reverse_components("foo/../bar/../mod.rs").collect();
+        assert_eq!(components, &["mod.rs"]);
+
+        let components: Vec<_> = reverse_components("foo/bar/foobar/../../mod.rs").collect();
+        assert_eq!(components, &["mod.rs", "foo"]);
+    }
+}
diff --git a/src/filetree/mod.rs b/src/filetree/mod.rs
new file mode 100644
index 0000000..4225604
--- /dev/null
+++ b/src/filetree/mod.rs
@@ -0,0 +1,130 @@
+use std::collections::BTreeMap;
+
+use components::reverse_components;
+
+mod components;
+
+#[derive(Default, Debug)]
+struct Node {
+    full_paths: Vec<String>,
+    children: BTreeMap<String, Node>,
+}
+
+#[derive(Default, Debug)]
+pub struct ReverseFileTree {
+    root: Node,
+}
+
+impl ReverseFileTree {
+    pub fn new() -> Self {
+        Self::default()
+    }
+
+    pub fn insert(&mut self, path: &str) {
+        let mut node = &mut self.root;
+
+        for component in reverse_components(path) {
+            node = node.children.entry(component.into()).or_default();
+        }
+
+        node.full_paths.push(path.into());
+    }
+
+    fn lookup(&self, path: &str, min_matches: Option<usize>) -> Vec<String> {
+        let mut matching_components = 0;
+        let mut components = reverse_components(path);
+        let mut node = &self.root;
+        let mut last_matching_paths = &vec![];
+
+        for component in &mut components {
+            match node.children.get(component) {
+                Some(child) => {
+                    matching_components += 1;
+                    node = child;
+                    if !node.full_paths.is_empty() {
+                        last_matching_paths = &node.full_paths;
+                    }
+                }
+                None => break,
+            }
+        }
+
+        let mut results = last_matching_paths.clone();
+        if matching_components >= min_matches.map_or(1, |n| n + 1) {
+            // we have exhausted all the path components, but the tree might still have more children
+            // so we follow a straight branch down the tree if one exists, and extend the results with whatever we find
+            while node.children.len() == 1 {
+                node = node.children.first_key_value().unwrap().1;
+                if !node.full_paths.is_empty() {
+                    results.extend_from_slice(&node.full_paths);
+                    return results;
+                }
+            }
+        }
+
+        results
+    }
+}
+
+impl<T> FromIterator<T> for ReverseFileTree
+where
+    T: AsRef<str>,
+{
+    fn from_iter<I: IntoIterator<Item = T>>(iter: I) -> Self {
+        let mut tree = Self::new();
+        for path in iter {
+            tree.insert(path.as_ref());
+        }
+        tree
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_min_matches() {
+        let tree = ReverseFileTree::from_iter(&["x/y/z"]);
+
+        let cases: &[(usize, &str, &[&str])] = &[
+            // only the basename has to match
+            (0, "z", &["x/y/z"]),
+            (0, "R/z", &["x/y/z"]),
+            (0, "R/y/z", &["x/y/z"]),
+            (0, "x/y/z", &["x/y/z"]),
+            (0, "w/x/y/z", &["x/y/z"]),
+            // basename + one ancestor have to match
+            (1, "z", &[]),
+            (1, "R/z", &[]),
+            (1, "R/y/z", &["x/y/z"]),
+            (1, "x/y/z", &["x/y/z"]),
+            (1, "w/x/y/z", &["x/y/z"]),
+            // 3 components have to match
+            (2, "z", &[]),
+            (2, "R/z", &[]),
+            (2, "R/y/z", &[]),
+            (2, "x/y/z", &["x/y/z"]),
+            (2, "w/x/y/z", &["x/y/z"]),
+        ];
+        for &(min_matches, lookup, result) in cases {
+            assert_eq!(tree.lookup(lookup, Some(min_matches)), result);
+        }
+    }
+
+    #[test]
+    fn test_lookup() {
+        let tree = ReverseFileTree::from_iter(&["mod.rs"]);
+        // exact lookup
+        assert_eq!(tree.lookup("mod.rs", None), &["mod.rs"]);
+        // no match
+        assert!(tree.lookup("not-found", None).is_empty());
+
+        let tree = ReverseFileTree::from_iter(&["foo/bar/mod.rs"]);
+
+        // the tree will follow unambiguous paths:
+        assert_eq!(tree.lookup("bar/mod.rs", None), &["foo/bar/mod.rs"]);
+        // it will also follow unambiguous partial matches:
+        assert_eq!(tree.lookup("qux/baz/bar/mod.rs", None), &["foo/bar/mod.rs"]);
+    }
+}
diff --git a/src/lib.rs b/src/lib.rs
index 6b0fc21..a90e602 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -3,6 +3,7 @@ use pyo3::prelude::*;
 
 mod compute_name;
 mod failure_message;
+mod filetree;
 mod junit;
 mod testrun;