feat(joblib): add

README.md

@@ -44,6 +44,7 @@ rw.write(df, "data.parquet")
 | [Binary](./readwrite/handlers/binary.py) | `bin` | [Python's `bytes`](https://docs.python.org/3/library/stdtypes.html#bytes) | :heavy_check_mark: | :heavy_check_mark: |
 | [Csv](./readwrite/handlers/csv.py) | `csv` | [`pandas`](https://pandas.pydata.org/) | :heavy_check_mark: | :heavy_check_mark: |
 | [Excel](./readwrite/handlers/excel.py) | `xlsx` | [`pandas`](https://pandas.pydata.org/) | :heavy_check_mark: | :heavy_check_mark: |
+| [Joblib](./readwrite/handlers/joblib.py) | `joblib` | [`joblib`](https://joblib.readthedocs.io/) | :heavy_check_mark: | :heavy_check_mark: |
 | [Json](./readwrite/handlers/json.py) | `json` | [Python's `json`](https://docs.python.org/3/library/json.html) | :heavy_check_mark: | :heavy_check_mark: |
 | [Parquet](./readwrite/handlers/parquet.py) | `parquet` | [`pandas`](https://pandas.pydata.org/) | :heavy_check_mark: | :heavy_check_mark: |
 | [Pickle](./readwrite/handlers/pickle.py) | `pkl`, `pickle` | [Python's `pickle`](https://docs.python.org/3/library/pickle.html) or [`pandas`](https://pandas.pydata.org/) | :heavy_check_mark: | :heavy_check_mark: |

readwrite/handlers/joblib.py

@@ -0,0 +1,20 @@
+from .base import Handler
+
+
+class JoblibHandler(Handler):
+
+    def __init__(self):
+        super().__init__(
+            "joblib",
+            ["joblib"]
+        )
+
+    def read(self, path: str, **kwargs):
+        import joblib
+
+        return joblib.load(path, **kwargs)
+
+    def write(self, x, path, **kwargs):
+        import joblib
+
+        joblib.dump(x, path, **kwargs)

readwrite/registry.py

@@ -1,10 +1,10 @@
 import os
 import typing
 
-from .constants import LOGGER
 from .handlers.base import Handler
 from .utils import measure_duration
 
+
 class UnknownExtension(ValueError):
     def __init__(self, extension: str):
         super().__init__(f"unknown extension: {extension}")
@@ -69,6 +69,9 @@ def add_defaults(self):
         from .handlers.excel import ExcelHandler
         self.add(ExcelHandler())
 
+        from .handlers.joblib import JoblibHandler
+        self.add(JoblibHandler())
+
         from .handlers.json import JsonHandler
         self.add(JsonHandler())
 

requirements.txt

@@ -7,3 +7,4 @@ tqdm
 pyyaml
 toml
 openpyxl
+joblib

tests/test_handlers.py

@@ -3,6 +3,7 @@
 import pickle
 import unittest
 
+import joblib
 import pandas
 import toml
 import yaml
@@ -59,6 +60,24 @@ def test_write(self):
         self.assertTrue(self.content.equals(pandas.read_csv(path)))
 
 
+class HandlerJoblibTest(unittest.TestCase):
+
+    handler = registry.get("joblib")
+    content = {'hello': 'world', 'from': 42}
+
+    def test_read(self):
+        path = fixture_path("hello.joblib")
+        x = self.handler.read(path)
+
+        self.assertEqual(self.content, x)
+
+    def test_write(self):
+        path = "/tmp/hello.joblib"
+        self.handler.write(self.content, path)
+
+        self.assertEqual(self.content, joblib.load(path))
+
+
 class HandlerExcelTest(unittest.TestCase):
 
     handler = registry.get("xlsx")