Partial JSON parsing support trailing strings (#101)

.github/workflows/ci.yml

@@ -57,7 +57,7 @@ jobs:
     - run: cargo careful t -F python
       if: matrix.rust-version == 'nightly'
 
-    - uses: codecov/codecov-action@v3
+    - uses: codecov/codecov-action@v4
       with:
         env_vars: RUNS_ON,RUST_VERSION
         token: ${{ secrets.CODECOV_TOKEN }}
@@ -103,11 +103,53 @@ jobs:
 
     - run: cargo test --doc
 
-    - uses: codecov/codecov-action@v3
+    - uses: codecov/codecov-action@v4
       with:
         env_vars: RUNS_ON,RUST_VERSION
         token: ${{ secrets.CODECOV_TOKEN }}
 
+  test-python:
+    name: test jiter-python
+
+    runs-on: ubuntu-latest
+
+    env:
+      RUNS_ON: ubuntu-latest
+
+    steps:
+    - uses: actions/checkout@v3
+
+    - name: set up python
+      uses: actions/setup-python@v4
+      with:
+        python-version: '3.12'
+
+    - uses: dtolnay/rust-toolchain@stable
+
+    - id: cache-rust
+      uses: Swatinem/rust-cache@v2
+
+    - run: cargo install rustfilt coverage-prepare
+      if: steps.cache-rust.outputs.cache-hit != 'true'
+
+    - run: rustup component add llvm-tools-preview
+
+    - run: make python-install
+
+    - run: pip install -e crates/jiter-python
+      env:
+        RUSTFLAGS: '-C instrument-coverage'
+
+    - run: pytest crates/jiter-python/tests
+      env:
+        RUST_BACKTRACE: 1
+
+    - run: coverage-prepare lcov $(python -c 'import jiter.jiter;print(jiter.jiter.__file__)')
+
+    - uses: codecov/codecov-action@v4
+      with:
+        token: ${{ secrets.CODECOV_TOKEN }}
+
   bench:
     runs-on: ubuntu-latest
     steps:
@@ -504,7 +546,7 @@ jobs:
   # https://github.com/marketplace/actions/alls-green#why used for branch protection checks
   check:
     if: always()
-    needs: [test-linux, test-macos, bench, fuzz, fuzz-skip, lint]
+    needs: [test-linux, test-macos, test-python, bench, fuzz, fuzz-skip, lint]
     runs-on: ubuntu-latest
     steps:
       - name: Decide whether the needed jobs succeeded or failed

Makefile

@@ -0,0 +1,44 @@
+.DEFAULT_GOAL := all
+
+.PHONY: format
+format:
+	@cargo fmt --version
+	cargo fmt
+
+.PHONY: lint
+lint:
+	@cargo clippy --version
+	cargo clippy -- -D warnings
+	cargo doc
+
+.PHONY: test
+test:
+	cargo test
+
+.PHONY: python-install
+python-install:
+	pip install maturin
+	pip install -r crates/jiter-python/tests/requirements.txt
+
+.PHONY: python-dev
+python-dev:
+	maturin develop -m crates/jiter-python/Cargo.toml
+
+.PHONY: python-test
+python-test: python-dev
+	pytest crates/jiter-python/tests
+
+.PHONY: bench
+bench:
+	cargo bench  -p jiter -F python
+
+.PHONY: fuzz
+fuzz:
+	cargo +nightly fuzz run --fuzz-dir crates/fuzz compare_to_serde --release
+
+.PHONY: fuzz-skip
+fuzz-skip:
+	cargo +nightly fuzz run --fuzz-dir crates/fuzz compare_skip --release
+
+.PHONY: all
+all: format lint test test-python

README.md

@@ -9,13 +9,13 @@ Fast iterable JSON parser.
 Documentation is available at [docs.rs/jiter](https://docs.rs/jiter).
 
 jiter has three interfaces:
-* [`JsonValue`] an enum representing JSON data
-* [`Jiter`] an iterator over JSON data
-* [`python_parse`] which parses a JSON string into a Python object
+* `JsonValue` an enum representing JSON data
+* `Jiter` an iterator over JSON data
+* `PythonParse` which parses a JSON string into a Python object
 
 ## JsonValue Example
 
-See [the `JsonValue` docs][JsonValue] for more details.
+See [the `JsonValue` docs](https://docs.rs/jiter/latest/jiter/enum.JsonValue.html) for more details.
 
 ```rust
 use jiter::JsonValue;
@@ -54,7 +54,7 @@ Object(
 
 ## Jiter Example
 
-To use [Jiter], you need to know what schema you're expecting:
+To use [Jiter](https://docs.rs/jiter/latest/jiter/struct.Jiter.html), you need to know what schema you're expecting:
 
 ```rust
 use jiter::{Jiter, NumberInt, Peek};
@@ -69,7 +69,7 @@ fn main() {
                 "+44 2345678"
             ]
         }"#;
-    let mut jiter = Jiter::new(json_data.as_bytes(), true);
+    let mut jiter = Jiter::new(json_data.as_bytes());
     assert_eq!(jiter.next_object().unwrap(), Some("name"));
     assert_eq!(jiter.next_str().unwrap(), "John Doe");
     assert_eq!(jiter.next_key().unwrap(), Some("age"));

crates/fuzz/fuzz_targets/compare_skip.rs

@@ -16,12 +16,12 @@ fuzz_target!(|json: String| {
     let json_data = json.as_bytes();
     match JsonValue::parse(json_data, false) {
         Ok(_) => {
-            let mut jiter = Jiter::new(json_data, false);
+            let mut jiter = Jiter::new(json_data);
             jiter.next_skip().unwrap();
             jiter.finish().unwrap();
         }
         Err(json_error) => {
-            let mut jiter = Jiter::new(json_data, false);
+            let mut jiter = Jiter::new(json_data);
             let jiter_error = match jiter.next_skip() {
                 Ok(_) => jiter.finish().unwrap_err(),
                 Err(e) => e,

crates/jiter-python/README.md

@@ -15,9 +15,10 @@ def from_json(
     /,
     *,
     allow_inf_nan: bool = True,
-    cache_strings: Literal[True, False, "all", "keys", "none"] = True,
-    allow_partial: bool = False,
+    cache_mode: Literal[True, False, "all", "keys", "none"] = "all",
+    partial_mode: Literal[True, False, "off", "on", "trailing-strings"] = False,
     catch_duplicate_keys: bool = False,
+    lossless_floats: bool = False,
 ) -> Any:
     """
     Parse input bytes into a JSON object.
@@ -26,12 +27,16 @@ def from_json(
         json_data: The JSON data to parse
         allow_inf_nan: Whether to allow infinity (`Infinity` an `-Infinity`) and `NaN` values to float fields.
             Defaults to True.
-        cache_strings: cache Python strings to improve performance at the cost of some memory usage
+        cache_mode: cache Python strings to improve performance at the cost of some memory usage
             - True / 'all' - cache all strings
             - 'keys' - cache only object keys
             - False / 'none' - cache nothing
-        allow_partial: if True, return parsed content when reaching EOF without closing objects and arrays
+        partial_mode: How to handle incomplete strings:
+            - False / 'off' - raise an exception if the input is incomplete
+            - True / 'on' - allow incomplete JSON but discard the last string if it is incomplete
+            - 'trailing-strings' - allow incomplete JSON, and include the last incomplete string in the output
         catch_duplicate_keys: if True, raise an exception if objects contain the same key multiple times
+        lossless_floats: if True, preserve full detail on floats using `LosslessFloat`
 
     Returns:
         Python object built from the JSON input.

crates/jiter-python/jiter.pyi

@@ -6,8 +6,8 @@ def from_json(
     /,
     *,
     allow_inf_nan: bool = True,
-    cache_strings: Literal[True, False, "all", "keys", "none"] = "all",
-    allow_partial: bool = False,
+    cache_mode: Literal[True, False, "all", "keys", "none"] = "all",
+    partial_mode: Literal[True, False, "off", "on", "trailing-strings"] = False,
     catch_duplicate_keys: bool = False,
     lossless_floats: bool = False,
 ) -> Any:
@@ -18,11 +18,14 @@ def from_json(
         json_data: The JSON data to parse
         allow_inf_nan: Whether to allow infinity (`Infinity` an `-Infinity`) and `NaN` values to float fields.
             Defaults to True.
-        cache_strings: cache Python strings to improve performance at the cost of some memory usage
+        cache_mode: cache Python strings to improve performance at the cost of some memory usage
             - True / 'all' - cache all strings
             - 'keys' - cache only object keys
             - False / 'none' - cache nothing
-        allow_partial: if True, return parsed content when reaching EOF without closing objects and arrays
+        partial_mode: How to handle incomplete strings:
+            - False / 'off' - raise an exception if the input is incomplete
+            - True / 'on' - allow incomplete JSON but discard the last string if it is incomplete
+            - 'trailing-strings' - allow incomplete JSON, and include the last incomplete string in the output
         catch_duplicate_keys: if True, raise an exception if objects contain the same key multiple times
         lossless_floats: if True, preserve full detail on floats using `LosslessFloat`
 

crates/jiter-python/src/lib.rs

@@ -2,7 +2,7 @@ use std::sync::OnceLock;
 
 use pyo3::prelude::*;
 
-use jiter::{map_json_error, LosslessFloat, PythonParseBuilder, StringCacheMode};
+use jiter::{map_json_error, LosslessFloat, PartialMode, PythonParse, StringCacheMode};
 
 #[allow(clippy::fn_params_excessive_bools)]
 #[pyfunction(
@@ -11,8 +11,8 @@ use jiter::{map_json_error, LosslessFloat, PythonParseBuilder, StringCacheMode};
         /,
         *,
         allow_inf_nan=true,
-        cache_strings=StringCacheMode::All,
-        allow_partial=false,
+        cache_mode=StringCacheMode::All,
+        partial_mode=PartialMode::Off,
         catch_duplicate_keys=false,
         lossless_floats=false,
     )
@@ -21,15 +21,15 @@ pub fn from_json<'py>(
     py: Python<'py>,
     json_data: &[u8],
     allow_inf_nan: bool,
-    cache_strings: StringCacheMode,
-    allow_partial: bool,
+    cache_mode: StringCacheMode,
+    partial_mode: PartialMode,
     catch_duplicate_keys: bool,
     lossless_floats: bool,
 ) -> PyResult<Bound<'py, PyAny>> {
-    let parse_builder = PythonParseBuilder {
+    let parse_builder = PythonParse {
         allow_inf_nan,
-        cache_mode: cache_strings,
-        allow_partial,
+        cache_mode,
+        partial_mode,
         catch_duplicate_keys,
         lossless_floats,
     };