From 7ef09cedb170a897e4e4277403025311a325dd0d Mon Sep 17 00:00:00 2001
From: Gary Benson <gary@gbenson.net>
Date: Fri, 31 May 2024 20:33:18 +0100
Subject: [PATCH] JSONL writer

---
 src/dom_tokenizers/internal/jsonl.py | 32 ++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)
 create mode 100644 src/dom_tokenizers/internal/jsonl.py

diff --git a/src/dom_tokenizers/internal/jsonl.py b/src/dom_tokenizers/internal/jsonl.py
new file mode 100644
index 0000000..e53908f
--- /dev/null
+++ b/src/dom_tokenizers/internal/jsonl.py
@@ -0,0 +1,32 @@
+from datetime import datetime
+from functools import cached_property
+from typing import Optional
+
+from . import json
+
+
+class Writer:
+    def __init__(
+            self,
+            filename: Optional[str] = None,
+            *,
+            mode: str = "a",
+            basename: Optional[str] = None,
+            ext: str = ".jsonl",
+            with_timestamp: bool = False,
+    ):
+        if filename is None:
+            filename = basename
+            if with_timestamp:
+                filename = f"{filename}-{datetime.now():%Y%m%d%H%M%S%f}"
+            filename = f"{filename}{ext}"
+        self.filename = filename
+        self._mode = mode
+
+    @cached_property
+    def _fp(self):
+        return open(self.filename, self._mode)
+
+    def write(self, **fields):
+        json.dump(fields, self._fp)
+        self._fp.write("\n")