Respect and enable uninstalls of existing .egg-info packages (#3380)

## Summary

Users often find themselves dropped into environments that contain
`.egg-info` packages. While we won't support installing these, it's not
hard to support identifying them (e.g., in `pip freeze`) and
_uninstalling_ them.

Closes #2841.
Closes #2928.
Closes #3341.

## Test Plan

Ran `cargo run pip freeze --python
/opt/homebrew/Caskroom/miniforge/base/envs/TEST/bin/python`, with an
environment that includes `pip` as an `.egg-info`
(`/opt/homebrew/Caskroom/miniforge/base/envs/TEST/lib/python3.12/site-packages/pip-24.0-py3.12.egg-info`):

```
cffi @ file:///Users/runner/miniforge3/conda-bld/cffi_1696001825047/work
pip==24.0
pycparser @ file:///home/conda/feedstock_root/build_artifacts/pycparser_1711811537435/work
setuptools==69.5.1
wheel==0.43.0
```

Then ran `cargo run pip uninstall`, verified that `pip` was uninstalled,
and no longer listed in `pip freeze`.

PIP_COMPATIBILITY.md

@@ -320,10 +320,12 @@ Unlike `pip`, uv does not enable keyring authentication by default.
 Unlike `pip`, uv does not wait until a request returns a HTTP 401 before searching for
 authentication. uv attaches authentication to all requests for hosts with credentials available.
 
-## Legacy features
+## `egg` support
 
 uv does not support features that are considered legacy or deprecated in `pip`. For example,
 uv does not support `.egg`-style distributions.
 
-uv does not plan to support features that the `pip` maintainers explicitly recommend against,
-like `--target`.
+However, uv does have partial support for `.egg-info`-style distributions, which are occasionally
+found in Docker images and Conda environments. Specifically, uv does not support installing new
+`.egg-info`-style distributions, but it will respect any existing `.egg-info`-style distributions
+during resolution, and can uninstall `.egg-info` distributions with `uv pip uninstall`.

crates/distribution-types/src/installed.rs

@@ -20,6 +20,8 @@ pub enum InstalledDist {
     Registry(InstalledRegistryDist),
     /// The distribution was derived from an arbitrary URL.
     Url(InstalledDirectUrlDist),
+    /// The distribution was derived from pre-existing `.egg-info` directory.
+    EggInfo(InstalledEggInfo),
 }
 
 #[derive(Debug, Clone)]
@@ -39,11 +41,26 @@ pub struct InstalledDirectUrlDist {
     pub path: PathBuf,
 }
 
+#[derive(Debug, Clone)]
+pub struct InstalledEggInfo {
+    pub name: PackageName,
+    pub version: Version,
+    pub path: PathBuf,
+}
+
+/// The format of the distribution.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub enum Format {
+    DistInfo,
+    EggInfo,
+}
+
 impl InstalledDist {
     /// Try to parse a distribution from a `.dist-info` directory name (like `django-5.0a1.dist-info`).
     ///
     /// See: <https://packaging.python.org/en/latest/specifications/recording-installed-packages/#recording-installed-packages>
     pub fn try_from_path(path: &Path) -> Result<Option<Self>> {
+        // Ex) `cffi-1.16.0.dist-info`
         if path.extension().is_some_and(|ext| ext == "dist-info") {
             let Some(file_stem) = path.file_stem() else {
                 return Ok(None);
@@ -84,14 +101,48 @@ impl InstalledDist {
                 })))
             };
         }
+
+        // Ex) `zstandard-0.22.0-py3.12.egg-info`
+        if path.extension().is_some_and(|ext| ext == "egg-info") {
+            let Some(file_stem) = path.file_stem() else {
+                return Ok(None);
+            };
+            let Some(file_stem) = file_stem.to_str() else {
+                return Ok(None);
+            };
+            let Some((name, version_python)) = file_stem.split_once('-') else {
+                return Ok(None);
+            };
+            let Some((version, _)) = version_python.split_once('-') else {
+                return Ok(None);
+            };
+            let name = PackageName::from_str(name)?;
+            let version = Version::from_str(version).map_err(|err| anyhow!(err))?;
+            return Ok(Some(Self::EggInfo(InstalledEggInfo {
+                name,
+                version,
+                path: path.to_path_buf(),
+            })));
+        }
+
         Ok(None)
     }
 
+    /// Return the [`Format`] of the distribution.
+    pub fn format(&self) -> Format {
+        match self {
+            Self::Registry(_) => Format::DistInfo,
+            Self::Url(_) => Format::DistInfo,
+            Self::EggInfo(_) => Format::EggInfo,
+        }
+    }
+
     /// Return the [`Path`] at which the distribution is stored on-disk.
     pub fn path(&self) -> &Path {
         match self {
             Self::Registry(dist) => &dist.path,
             Self::Url(dist) => &dist.path,
+            Self::EggInfo(dist) => &dist.path,
         }
     }
 
@@ -100,6 +151,7 @@ impl InstalledDist {
         match self {
             Self::Registry(dist) => &dist.version,
             Self::Url(dist) => &dist.version,
+            Self::EggInfo(dist) => &dist.version,
         }
     }
 
@@ -115,11 +167,29 @@ impl InstalledDist {
 
     /// Read the `METADATA` file from a `.dist-info` directory.
     pub fn metadata(&self) -> Result<pypi_types::Metadata23> {
-        let path = self.path().join("METADATA");
-        let contents = fs::read(&path)?;
-        // TODO(zanieb): Update this to use thiserror so we can unpack parse errors downstream
-        pypi_types::Metadata23::parse_metadata(&contents)
-            .with_context(|| format!("Failed to parse METADATA file at: {}", path.user_display()))
+        match self.format() {
+            Format::DistInfo => {
+                let path = self.path().join("METADATA");
+                let contents = fs::read(&path)?;
+                // TODO(zanieb): Update this to use thiserror so we can unpack parse errors downstream
+                pypi_types::Metadata23::parse_metadata(&contents).with_context(|| {
+                    format!(
+                        "Failed to parse `METADATA` file at: {}",
+                        path.user_display()
+                    )
+                })
+            }
+            Format::EggInfo => {
+                let path = self.path().join("PKG-INFO");
+                let contents = fs::read(&path)?;
+                pypi_types::Metadata23::parse_metadata(&contents).with_context(|| {
+                    format!(
+                        "Failed to parse `PKG-INFO` file at: {}",
+                        path.user_display()
+                    )
+                })
+            }
+        }
     }
 
     /// Return the `INSTALLER` of the distribution.
@@ -137,6 +207,7 @@ impl InstalledDist {
         match self {
             Self::Registry(_) => false,
             Self::Url(dist) => dist.editable,
+            Self::EggInfo(_) => false,
         }
     }
 
@@ -145,6 +216,7 @@ impl InstalledDist {
         match self {
             Self::Registry(_) => None,
             Self::Url(dist) => dist.editable.then_some(&dist.url),
+            Self::EggInfo(_) => None,
         }
     }
 }
@@ -167,11 +239,18 @@ impl Name for InstalledDirectUrlDist {
     }
 }
 
+impl Name for InstalledEggInfo {
+    fn name(&self) -> &PackageName {
+        &self.name
+    }
+}
+
 impl Name for InstalledDist {
     fn name(&self) -> &PackageName {
         match self {
             Self::Registry(dist) => dist.name(),
             Self::Url(dist) => dist.name(),
+            Self::EggInfo(dist) => dist.name(),
         }
     }
 }
@@ -188,11 +267,18 @@ impl InstalledMetadata for InstalledDirectUrlDist {
     }
 }
 
+impl InstalledMetadata for InstalledEggInfo {
+    fn installed_version(&self) -> InstalledVersion {
+        InstalledVersion::Version(&self.version)
+    }
+}
+
 impl InstalledMetadata for InstalledDist {
     fn installed_version(&self) -> InstalledVersion {
         match self {
             Self::Registry(dist) => dist.installed_version(),
             Self::Url(dist) => dist.installed_version(),
+            Self::EggInfo(dist) => dist.installed_version(),
         }
     }
 }

crates/install-wheel-rs/src/lib.rs

@@ -11,7 +11,7 @@ use zip::result::ZipError;
 use pep440_rs::Version;
 use platform_tags::{Arch, Os};
 use pypi_types::Scheme;
-pub use uninstall::{uninstall_wheel, Uninstall};
+pub use uninstall::{uninstall_egg, uninstall_wheel, Uninstall};
 use uv_fs::Simplified;
 use uv_normalize::PackageName;
 
@@ -82,8 +82,10 @@ pub enum Error {
     DirectUrlJson(#[from] serde_json::Error),
     #[error("No .dist-info directory found")]
     MissingDistInfo,
-    #[error("Cannot uninstall package; RECORD file not found at: {}", _0.user_display())]
+    #[error("Cannot uninstall package; `RECORD` file not found at: {}", _0.user_display())]
     MissingRecord(PathBuf),
+    #[error("Cannot uninstall package; `top_level.txt` file not found at: {}", _0.user_display())]
+    MissingTopLevel(PathBuf),
     #[error("Multiple .dist-info directories found: {0}")]
     MultipleDistInfo(String),
     #[error(

crates/install-wheel-rs/src/uninstall.rs

@@ -7,7 +7,7 @@ use tracing::debug;
 use crate::wheel::read_record_file;
 use crate::Error;
 
-/// Uninstall the wheel represented by the given `dist_info` directory.
+/// Uninstall the wheel represented by the given `.dist-info` directory.
 pub fn uninstall_wheel(dist_info: &Path) -> Result<Uninstall, Error> {
     let Some(site_packages) = dist_info.parent() else {
         return Err(Error::BrokenVenv(
@@ -118,6 +118,96 @@ pub fn uninstall_wheel(dist_info: &Path) -> Result<Uninstall, Error> {
     })
 }
 
+/// Uninstall the egg represented by the `.egg-info` directory.
+///
+/// See: <https://github.com/pypa/pip/blob/41587f5e0017bcd849f42b314dc8a34a7db75621/src/pip/_internal/req/req_uninstall.py#L483>
+pub fn uninstall_egg(egg_info: &Path) -> Result<Uninstall, Error> {
+    let mut file_count = 0usize;
+    let mut dir_count = 0usize;
+
+    let dist_location = egg_info
+        .parent()
+        .expect("egg-info directory is not in a site-packages directory");
+
+    // Read the `namespace_packages.txt` file.
+    let namespace_packages = {
+        let namespace_packages_path = egg_info.join("namespace_packages.txt");
+        match fs_err::read_to_string(namespace_packages_path) {
+            Ok(namespace_packages) => namespace_packages
+                .lines()
+                .map(ToString::to_string)
+                .collect::<Vec<_>>(),
+            Err(err) if err.kind() == std::io::ErrorKind::NotFound => {
+                vec![]
+            }
+            Err(err) => return Err(err.into()),
+        }
+    };
+
+    // Read the `top_level.txt` file, ignoring anything in `namespace_packages.txt`.
+    let top_level = {
+        let top_level_path = egg_info.join("top_level.txt");
+        match fs_err::read_to_string(&top_level_path) {
+            Ok(top_level) => top_level
+                .lines()
+                .map(ToString::to_string)
+                .filter(|line| !namespace_packages.contains(line))
+                .collect::<Vec<_>>(),
+            Err(err) if err.kind() == std::io::ErrorKind::NotFound => {
+                return Err(Error::MissingTopLevel(top_level_path));
+            }
+            Err(err) => return Err(err.into()),
+        }
+    };
+
+    // Remove everything in `top_level.txt`.
+    for entry in top_level {
+        let path = dist_location.join(&entry);
+
+        // Remove as a directory.
+        match fs_err::remove_dir_all(&path) {
+            Ok(()) => {
+                debug!("Removed directory: {}", path.display());
+                dir_count += 1;
+                continue;
+            }
+            Err(err) if err.kind() == std::io::ErrorKind::NotFound => {}
+            Err(err) => return Err(err.into()),
+        }
+
+        // Remove as a `.py`, `.pyc`, or `.pyo` file.
+        for exten in &["py", "pyc", "pyo"] {
+            let path = path.with_extension(exten);
+            match fs_err::remove_file(&path) {
+                Ok(()) => {
+                    debug!("Removed file: {}", path.display());
+                    file_count += 1;
+                    break;
+                }
+                Err(err) if err.kind() == std::io::ErrorKind::NotFound => {}
+                Err(err) => return Err(err.into()),
+            }
+        }
+    }
+
+    // Remove the `.egg-info` directory.
+    match fs_err::remove_dir_all(egg_info) {
+        Ok(()) => {
+            debug!("Removed directory: {}", egg_info.display());
+            dir_count += 1;
+        }
+        Err(err) if err.kind() == std::io::ErrorKind::NotFound => {}
+        Err(err) => {
+            return Err(err.into());
+        }
+    }
+
+    Ok(Uninstall {
+        file_count,
+        dir_count,
+    })
+}
+
 #[derive(Debug, Default)]
 pub struct Uninstall {
     /// The number of files that were removed during the uninstallation.

crates/uv-installer/src/uninstall.rs

@@ -1,14 +1,18 @@
 use anyhow::Result;
 
-use distribution_types::InstalledDist;
+use distribution_types::{Format, InstalledDist};
 
 /// Uninstall a package from the specified Python environment.
 pub async fn uninstall(
     dist: &InstalledDist,
 ) -> Result<install_wheel_rs::Uninstall, UninstallError> {
     let uninstall = tokio::task::spawn_blocking({
         let path = dist.path().to_owned();
-        move || install_wheel_rs::uninstall_wheel(&path)
+        let format = dist.format();
+        move || match format {
+            Format::DistInfo => install_wheel_rs::uninstall_wheel(&path),
+            Format::EggInfo => install_wheel_rs::uninstall_egg(&path),
+        }
     })
     .await??;
 

crates/uv/src/commands/pip_freeze.rs

@@ -62,6 +62,9 @@ pub(crate) fn pip_freeze(
                     writeln!(printer.stdout(), "{} @ {}", dist.name().bold(), dist.url)?;
                 }
             }
+            InstalledDist::EggInfo(dist) => {
+                writeln!(printer.stdout(), "{}=={}", dist.name().bold(), dist.version)?;
+            }
         }
     }