Merge remote-tracking branch 'origin/main' into fix-explode-copy

deutsche-nationalbibliothek · May 29, 2024 · b9973be · b9973be
2 parents 3bd81a8 + 3efe646
commit b9973be
Show file tree

Hide file tree

Showing 5 changed files with 34 additions and 34 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -33,15 +33,15 @@ jobs:
       matrix:
         target:
           - { os: ubuntu-latest, toolchain: stable, triple: x86_64-unknown-linux-gnu }
-          - { os: ubuntu-latest, toolchain: 1.74.1, triple: x86_64-unknown-linux-gnu }
+          - { os: ubuntu-latest, toolchain: 1.76, triple: x86_64-unknown-linux-gnu }
           - { os: ubuntu-latest, toolchain: beta, triple: x86_64-unknown-linux-gnu }
           - { os: ubuntu-latest, toolchain: nightly, triple: x86_64-unknown-linux-gnu }
           - { os: macos-latest, toolchain: stable, triple: x86_64-apple-darwin }
-          - { os: macos-latest, toolchain: 1.74.1, triple: x86_64-apple-darwin }
+          - { os: macos-latest, toolchain: 1.76, triple: x86_64-apple-darwin }
           - { os: windows-latest, toolchain: stable, triple: x86_64-pc-windows-gnu }
-          - { os: windows-latest, toolchain: 1.74.1, triple: x86_64-pc-windows-gnu }
+          - { os: windows-latest, toolchain: 1.76, triple: x86_64-pc-windows-gnu }
           - { os: windows-latest, toolchain: stable, triple: i686-pc-windows-msvc }
-          - { os: windows-latest, toolchain: 1.74.1, triple: i686-pc-windows-msvc }
+          - { os: windows-latest, toolchain: 1.76, triple: i686-pc-windows-msvc }
     runs-on: ${{ matrix.target.os }}
     steps:
       - uses: actions/checkout@v4
@@ -183,7 +183,8 @@ jobs:
       - uses: dtolnay/rust-toolchain@stable
       - uses: dtolnay/install@cargo-outdated
       - uses: Swatinem/rust-cache@v2
-      - run: cargo outdated --exit-code 1
+      - run: cargo update
+      - run: cargo outdated --workspace --exit-code 1
 
   miri:
     name: miri

diff --git a/.github/workflows/daily.yml b/.github/workflows/daily.yml
@@ -25,4 +25,5 @@ jobs:
       - uses: dtolnay/rust-toolchain@stable
       - uses: dtolnay/install@cargo-outdated
       - uses: Swatinem/rust-cache@v2
-      - run: cargo outdated --exit-code 1
+      - run: cargo update
+      - run: cargo outdated --workspace --exit-code 1
diff --git a/Cargo.toml b/Cargo.toml
@@ -19,7 +19,7 @@ license = "EUPL-1.2"
 readme = "./README.md"
 keywords = ["PICA+", "code4lib"]
 edition = "2021"
-rust-version = "1.74.1"
+rust-version = "1.76"
 
 [workspace.dependencies]
 pica-matcher = { version = "0.24", path = "./crates/pica-matcher" }
@@ -31,11 +31,11 @@ pica-utils = { version = "0.24", path = "./crates/pica-utils" }
 anyhow = "1.0"
 bstr = "1.9"
 chrono = { version = "0.4", default-features = false }
-clap = "4.4"
-clap_complete = "4.4"
+clap = "4.5"
+clap_complete = "4.5"
 csv = "1.3"
 flate2 = "1.0"
-polars = { version = "0.38", features = ["ipc", "decompress", "performant"] }
+polars = { version = "0.40", features = ["ipc", "decompress", "performant"] }
 quickcheck = "1.0"
 rand = "0.8"
 regex = "1.10"

diff --git a/crates/pica-toolkit/Cargo.toml b/crates/pica-toolkit/Cargo.toml
@@ -34,11 +34,8 @@ toml = { workspace = true }
 unicode-normalization = { version = "0.1" }
 
 [dev-dependencies]
-assert_cmd = "2.0"
-predicates = "3.0"
 quickcheck = "1.0"
 quickcheck_macros = "1.0"
-tempfile = "3.8"
 trycmd = "0.15"
 
 [[bin]]

diff --git a/crates/pica-toolkit/src/filter_list.rs b/crates/pica-toolkit/src/filter_list.rs
@@ -93,35 +93,36 @@ impl FilterList {
     ) -> Result<DataFrame, FilterListError> {
         let extension = path.extension().and_then(OsStr::to_str);
         let path_str = path.to_str().unwrap_or_default();
+        let path = path.to_owned();
+
+        let options = CsvReadOptions::default()
+            .with_has_header(true)
+            .with_infer_schema_length(Some(0));
 
         match extension {
             Some("ipc" | "arrow" | "feather") => {
                 Ok(IpcReader::new(File::open(path)?)
-                    .memory_mapped(false)
+                    .memory_mapped(None)
                     .finish()?)
             }
-            Some("csv") => Ok(CsvReader::from_path(path)?
-                .infer_schema(Some(0))
-                .has_header(true)
+            Some("csv") => Ok(options
+                .try_into_reader_with_file_path(Some(path))?
                 .finish()?),
-            Some("gz") if path_str.ends_with(".csv.gz") => {
-                Ok(CsvReader::from_path(path)?
-                    .infer_schema(Some(0))
-                    .has_header(true)
-                    .finish()?)
-            }
-            Some("tsv") => Ok(CsvReader::from_path(path)?
-                .with_separator(b'\t')
-                .has_header(true)
-                .infer_schema(Some(0))
+            Some("gz") if path_str.ends_with(".csv.gz") => Ok(options
+                .try_into_reader_with_file_path(Some(path))?
+                .finish()?),
+            Some("tsv") => Ok(options
+                .with_parse_options(
+                    CsvParseOptions::default().with_separator(b'\t'),
+                )
+                .try_into_reader_with_file_path(Some(path))?
+                .finish()?),
+            Some("gz") if path_str.ends_with(".tsv.gz") => Ok(options
+                .with_parse_options(
+                    CsvParseOptions::default().with_separator(b'\t'),
+                )
+                .try_into_reader_with_file_path(Some(path))?
                 .finish()?),
-            Some("gz") if path_str.ends_with(".tsv.gz") => {
-                Ok(CsvReader::from_path(path)?
-                    .with_separator(b'\t')
-                    .infer_schema(Some(0))
-                    .has_header(true)
-                    .finish()?)
-            }
             _ => {
                 Err(FilterListError::InvalidFileFormat(path_str.into()))
             }