starred/minne
1
0
Fork 0
mirror of https://github.com/perstarkse/minne.git synced 2026-05-30 03:10:45 +02:00
Code Issues Packages Projects Releases 10 Wiki Activity

clippy: evaluations crate

Browse Source
This commit is contained in:
Per Stark
2026-05-26 20:07:51 +02:00
parent b4383bb227
commit 041d9bd81f
16 changed files with 88 additions and 118 deletions
Download Patch File Download Diff File Expand all files Collapse all files
evaluations/src/datasets/beir.rs
+21 -30
View File
@@ -59,7 +59,7 @@ pub fn convert_beir(raw_dir: &Path, dataset: DatasetKind) -> Result<Vec<Converte
let mut paragraphs = Vec::with_capacity(corpus.len());
let mut paragraph_index = HashMap::new();
for (doc_id, entry) in corpus.iter() {
for (doc_id, entry) in &corpus {
let paragraph_id = format!("{}-{doc_id}", dataset.source_prefix());
let paragraph = ConvertedParagraph {
id: paragraph_id.clone(),
@@ -76,13 +76,10 @@ pub fn convert_beir(raw_dir: &Path, dataset: DatasetKind) -> Result<Vec<Converte
let mut skipped_answers = 0usize;
for (query_id, entries) in qrels {
let query = match queries.get(&query_id) {
Some(query) => query,
None => {
missing_queries += 1;
warn!(query_id = %query_id, "Skipping qrels entry for missing query");
continue;
}
let query = if let Some(query) = queries.get(&query_id) { query } else {
missing_queries += 1;
warn!(query_id = %query_id, "Skipping qrels entry for missing query");
continue;
};
let best = match select_best_doc(&entries) {
@@ -90,31 +87,25 @@ pub fn convert_beir(raw_dir: &Path, dataset: DatasetKind) -> Result<Vec<Converte
None => continue,
};
let paragraph_slot = match paragraph_index.get(&best.doc_id) {
Some(slot) => *slot,
None => {
missing_docs += 1;
warn!(
query_id = %query_id,
doc_id = %best.doc_id,
"Skipping qrels entry referencing missing corpus document"
);
continue;
}
let paragraph_slot = if let Some(slot) = paragraph_index.get(&best.doc_id) { *slot } else {
missing_docs += 1;
warn!(
query_id = %query_id,
doc_id = %best.doc_id,
"Skipping qrels entry referencing missing corpus document"
);
continue;
};
let answer = answer_snippet(&paragraphs[paragraph_slot].context);
let answers = match answer {
Some(snippet) => vec![snippet],
None => {
skipped_answers += 1;
warn!(
query_id = %query_id,
doc_id = %best.doc_id,
"Skipping query because no non-empty answer snippet could be derived"
);
continue;
}
let answers = if let Some(snippet) = answer { vec![snippet] } else {
skipped_answers += 1;
warn!(
query_id = %query_id,
doc_id = %best.doc_id,
"Skipping query because no non-empty answer snippet could be derived"
);
continue;
};
let question_id = format!("{}-{query_id}", dataset.source_prefix());
evaluations/src/datasets/mod.rs
+4 -8
View File
@@ -99,9 +99,9 @@ struct ManifestSlice {
impl DatasetCatalog {
pub fn load() -> Result<Self> {
let manifest_raw = fs::read_to_string(MANIFEST_PATH)
.with_context(|| format!("reading dataset manifest at {}", MANIFEST_PATH))?;
.with_context(|| format!("reading dataset manifest at {MANIFEST_PATH}"))?;
let manifest: ManifestFile = serde_yaml::from_str(&manifest_raw)
.with_context(|| format!("parsing dataset manifest at {}", MANIFEST_PATH))?;
.with_context(|| format!("parsing dataset manifest at {MANIFEST_PATH}"))?;
let root = Path::new(env!("CARGO_MANIFEST_DIR"));
let mut datasets = BTreeMap::new();
@@ -351,15 +351,11 @@ impl DatasetKind {
}
pub fn default_raw_path(self) -> PathBuf {
dataset_entry_for_kind(self)
.map(|entry| entry.raw_path.clone())
.unwrap_or_else(|err| panic!("dataset manifest missing entry for {:?}: {err}", self))
dataset_entry_for_kind(self).map_or_else(|err| panic!("dataset manifest missing entry for {self:?}: {err}"), |entry| entry.raw_path.clone())
}
pub fn default_converted_path(self) -> PathBuf {
dataset_entry_for_kind(self)
.map(|entry| entry.converted_path.clone())
.unwrap_or_else(|err| panic!("dataset manifest missing entry for {:?}: {err}", self))
dataset_entry_for_kind(self).map_or_else(|err| panic!("dataset manifest missing entry for {self:?}: {err}"), |entry| entry.converted_path.clone())
}
}