meilisearch-python-sdk 3.3.0__py3-none-any.whl → 3.5.0__py3-none-any.whl

meilisearch_python_sdk/_client.py

@@ -150,6 +150,7 @@ class AsyncClient(BaseClient):
         verify: str | bool | SSLContext = True,
         custom_headers: dict[str, str] | None = None,
         json_handler: BuiltinHandler | OrjsonHandler | UjsonHandler | None = None,
+        http2: bool = False,
     ) -> None:
         """Class initializer.
 
@@ -168,11 +169,12 @@ class AsyncClient(BaseClient):
                 (uses the json module from the standard library), OrjsonHandler (uses orjson), or
                 UjsonHandler (uses ujson). Note that in order use orjson or ujson the corresponding
                 extra needs to be included. Default: BuiltinHandler.
+            http2: Whether or not to use HTTP/2. Defaults to False.
         """
         super().__init__(api_key, custom_headers, json_handler)
 
         self.http_client = HttpxAsyncClient(
-            base_url=url, timeout=timeout, headers=self._headers, verify=verify
+            base_url=url, timeout=timeout, headers=self._headers, verify=verify, http2=http2
         )
         self._http_requests = AsyncHttpRequests(self.http_client, json_handler=self.json_handler)
 
@@ -649,8 +651,9 @@ class AsyncClient(BaseClient):
         Args:
 
             queries: List of SearchParameters
-            federation: If included a single search result with hits built from all queries. This
-                parameter can only be used with Meilisearch >= v1.10.0. Defaults to None.
+            federation: If included a single search result with hits built from all queries will
+                be returned. This parameter can only be used with Meilisearch >= v1.10.0. Defaults
+                to None.
             hits_type: Allows for a custom type to be passed to use for hits. Defaults to
                 JsonDict
 
@@ -1072,6 +1075,7 @@ class Client(BaseClient):
         verify: str | bool | SSLContext = True,
         custom_headers: dict[str, str] | None = None,
         json_handler: BuiltinHandler | OrjsonHandler | UjsonHandler | None = None,
+        http2: bool = False,
     ) -> None:
         """Class initializer.
 
@@ -1090,12 +1094,14 @@ class Client(BaseClient):
                 (uses the json module from the standard library), OrjsonHandler (uses orjson), or
                 UjsonHandler (uses ujson). Note that in order use orjson or ujson the corresponding
                 extra needs to be included. Default: BuiltinHandler.
+            http2: If set to True, the client will use HTTP/2. Defaults to False.
         """
         super().__init__(api_key, custom_headers, json_handler)
 
         self.http_client = HttpxClient(
-            base_url=url, timeout=timeout, headers=self._headers, verify=verify
+            base_url=url, timeout=timeout, headers=self._headers, verify=verify, http2=http2
         )
+
         self._http_requests = HttpRequests(self.http_client, json_handler=self.json_handler)
 
     def create_dump(self) -> TaskInfo:
@@ -1549,8 +1555,9 @@ class Client(BaseClient):
         Args:
 
             queries: List of SearchParameters
-            federation: If included a single search result with hits built from all queries. This
-                parameter can only be used with Meilisearch >= v1.10.0. Defaults to None.
+            federation: If included a single search result with hits built from all queries will
+                be returned. This parameter can only be used with Meilisearch >= v1.10.0. Defaults
+                to None.
             hits_type: Allows for a custom type to be passed to use for hits. Defaults to
                 JsonDict
 

meilisearch_python_sdk/_version.py

@@ -1 +1 @@
-VERSION = "3.3.0"
+VERSION = "3.5.0"

meilisearch_python_sdk/decorators.py

@@ -26,6 +26,7 @@ def async_add_documents(
     batch_size: int | None = None,
     primary_key: str | None = None,
     wait_for_task: bool = False,
+    verify: bool = True,
 ) -> Callable:
     """Decorator that takes the returned documents from a function and asyncronously adds them to Meilisearch.
 
@@ -42,6 +43,7 @@ def async_add_documents(
             Defaults to None.
         wait_for_task: If set to `True` the decorator will wait for the document addition to finish
             indexing before returning, otherwise it will return right away. Default = False.
+        verify: If set to `False` the decorator will not verify the SSL certificate of the server.
 
     Returns:
 
@@ -89,7 +91,9 @@ def async_add_documents(
                 )
                 return result
 
-            async with AsyncClient(connection_info.url, connection_info.api_key) as client:
+            async with AsyncClient(
+                connection_info.url, connection_info.api_key, verify=verify
+            ) as client:
                 await _async_add_documents(
                     client, index_name, result, batch_size, primary_key, wait_for_task
                 )
@@ -108,6 +112,7 @@ def add_documents(
     batch_size: int | None = None,
     primary_key: str | None = None,
     wait_for_task: bool = False,
+    verify: bool = True,
 ) -> Callable:
     """Decorator that takes the returned documents from a function and adds them to Meilisearch.
 
@@ -124,6 +129,7 @@ def add_documents(
             Defaults to None.
         wait_for_task: If set to `True` the decorator will wait for the document addition to finish
             indexing before returning, otherwise it will return right away. Default = False.
+        verify: If set to `False` the decorator will not verify the SSL certificate of the server.
 
     Returns:
 
@@ -171,7 +177,9 @@ def add_documents(
                 )
                 return result
 
-            decorator_client = Client(url=connection_info.url, api_key=connection_info.api_key)
+            decorator_client = Client(
+                url=connection_info.url, api_key=connection_info.api_key, verify=verify
+            )
             _add_documents(
                 decorator_client,
                 index_name,

meilisearch_python_sdk/index.py

@@ -1596,6 +1596,7 @@ class AsyncIndex(_BaseIndex):
         batch_size: int = 1000,
         primary_key: str | None = None,
         compress: bool = False,
+        concurrency_limit: int | None = None,
     ) -> list[TaskInfo]:
         """Adds documents in batches to reduce RAM usage with indexing.
 
@@ -1607,6 +1608,9 @@ class AsyncIndex(_BaseIndex):
             primary_key: The primary key of the documents. This will be ignored if already set.
                 Defaults to None.
             compress: If set to True the data will be sent in gzip format. Defaults to False.
+            concurrency_limit: If set this will limit the number of batches that will be sent
+                concurrently. This can be helpful if you find you are overloading the Meilisearch
+                server with requests. Defaults to None.
 
         Returns:
 
@@ -1628,6 +1632,23 @@ class AsyncIndex(_BaseIndex):
             >>>     index = client.index("movies")
             >>>     await index.add_documents_in_batches(documents)
         """
+        if concurrency_limit:
+            async with asyncio.Semaphore(concurrency_limit):
+                if not use_task_groups():
+                    batches = [
+                        self.add_documents(x, primary_key, compress=compress)
+                        for x in _batch(documents, batch_size)
+                    ]
+                    return await asyncio.gather(*batches)
+
+                async with asyncio.TaskGroup() as tg:  # type: ignore[attr-defined]
+                    tasks = [
+                        tg.create_task(self.add_documents(x, primary_key, compress=compress))
+                        for x in _batch(documents, batch_size)
+                    ]
+
+                return [x.result() for x in tasks]
+
         if not use_task_groups():
             batches = [
                 self.add_documents(x, primary_key, compress=compress)
@@ -1652,6 +1673,7 @@ class AsyncIndex(_BaseIndex):
         csv_delimiter: str | None = None,
         combine_documents: bool = True,
         compress: bool = False,
+        concurrency_limit: int | None = None,
     ) -> list[TaskInfo]:
         """Load all json files from a directory and add the documents to the index.
 
@@ -1668,6 +1690,9 @@ class AsyncIndex(_BaseIndex):
             combine_documents: If set to True this will combine the documents from all the files
                 before indexing them. Defaults to True.
             compress: If set to True the data will be sent in gzip format. Defaults to False.
+            concurrency_limit: If set this will limit the number of batches that will be sent
+                concurrently. This can be helpful if you find you are overloading the Meilisearch
+                server with requests. Defaults to None.
 
         Returns:
 
@@ -1709,6 +1734,54 @@ class AsyncIndex(_BaseIndex):
 
             return [response]
 
+        if concurrency_limit:
+            async with asyncio.Semaphore(concurrency_limit):
+                if not use_task_groups():
+                    add_documents = []
+                    for path in directory.iterdir():
+                        if path.suffix == f".{document_type}":
+                            documents = await _async_load_documents_from_file(
+                                path, csv_delimiter, json_handler=self._json_handler
+                            )
+                            add_documents.append(
+                                self.add_documents(documents, primary_key, compress=compress)
+                            )
+
+                    _raise_on_no_documents(add_documents, document_type, directory_path)
+
+                    if len(add_documents) > 1:
+                        # Send the first document on its own before starting the gather. Otherwise Meilisearch
+                        # returns an error because it thinks all entries are trying to create the same index.
+                        first_response = [await add_documents.pop()]
+
+                        responses = await asyncio.gather(*add_documents)
+                        responses = [*first_response, *responses]
+                    else:
+                        responses = [await add_documents[0]]
+
+                    return responses
+
+                async with asyncio.TaskGroup() as tg:  # type: ignore[attr-defined]
+                    tasks = []
+                    all_results = []
+                    for i, path in enumerate(directory.iterdir()):
+                        if path.suffix == f".{document_type}":
+                            documents = await _async_load_documents_from_file(
+                                path, csv_delimiter, json_handler=self._json_handler
+                            )
+                            if i == 0:
+                                all_results = [
+                                    await self.add_documents(documents, compress=compress)
+                                ]
+                            else:
+                                tasks.append(
+                                    tg.create_task(
+                                        self.add_documents(
+                                            documents, primary_key, compress=compress
+                                        )
+                                    )
+                                )
+
         if not use_task_groups():
             add_documents = []
             for path in directory.iterdir():
@@ -1766,6 +1839,7 @@ class AsyncIndex(_BaseIndex):
         csv_delimiter: str | None = None,
         combine_documents: bool = True,
         compress: bool = False,
+        concurrency_limit: int | None = None,
     ) -> list[TaskInfo]:
         """Load all json files from a directory and add the documents to the index in batches.
 
@@ -1784,6 +1858,9 @@ class AsyncIndex(_BaseIndex):
             combine_documents: If set to True this will combine the documents from all the files
                 before indexing them. Defaults to True.
             compress: If set to True the data will be sent in gzip format. Defaults to False.
+            concurrency_limit: If set this will limit the number of batches that will be sent
+                concurrently. This can be helpful if you find you are overloading the Meilisearch
+                server with requests. Defaults to None.
 
         Returns:
 
@@ -1826,6 +1903,7 @@ class AsyncIndex(_BaseIndex):
                 batch_size=batch_size,
                 primary_key=primary_key,
                 compress=compress,
+                concurrency_limit=concurrency_limit,
             )
 
         responses: list[TaskInfo] = []
@@ -1842,6 +1920,7 @@ class AsyncIndex(_BaseIndex):
                         batch_size=batch_size,
                         primary_key=primary_key,
                         compress=compress,
+                        concurrency_limit=concurrency_limit,
                     )
                 )
 
@@ -1908,6 +1987,7 @@ class AsyncIndex(_BaseIndex):
         primary_key: str | None = None,
         csv_delimiter: str | None = None,
         compress: bool = False,
+        concurrency_limit: int | None = None,
     ) -> list[TaskInfo]:
         """Adds documents form a json file in batches to reduce RAM usage with indexing.
 
@@ -1921,6 +2001,9 @@ class AsyncIndex(_BaseIndex):
             csv_delimiter: A single ASCII character to specify the delimiter for csv files. This
                 can only be used if the file is a csv file. Defaults to comma.
             compress: If set to True the data will be sent in gzip format. Defaults to False.
+            concurrency_limit: If set this will limit the number of batches that will be sent
+                concurrently. This can be helpful if you find you are overloading the Meilisearch
+                server with requests. Defaults to None.
 
         Returns:
 
@@ -1951,6 +2034,7 @@ class AsyncIndex(_BaseIndex):
             batch_size=batch_size,
             primary_key=primary_key,
             compress=compress,
+            concurrency_limit=concurrency_limit,
         )
 
     async def add_documents_from_raw_file(
@@ -2028,7 +2112,7 @@ class AsyncIndex(_BaseIndex):
         else:
             url = self._documents_url
 
-        async with aiofiles.open(upload_path, "r") as f:
+        async with aiofiles.open(upload_path) as f:
             data = await f.read()
 
         response = await self._http_requests.post(
@@ -2232,6 +2316,7 @@ class AsyncIndex(_BaseIndex):
         batch_size: int = 1000,
         primary_key: str | None = None,
         compress: bool = False,
+        concurrency_limit: int | None = None,
     ) -> list[TaskInfo]:
         """Update documents in batches to reduce RAM usage with indexing.
 
@@ -2245,6 +2330,9 @@ class AsyncIndex(_BaseIndex):
             primary_key: The primary key of the documents. This will be ignored if already set.
                 Defaults to None.
             compress: If set to True the data will be sent in gzip format. Defaults to False.
+            concurrency_limit: If set this will limit the number of batches that will be sent
+                concurrently. This can be helpful if you find you are overloading the Meilisearch
+                server with requests. Defaults to None.
 
         Returns:
 
@@ -2266,6 +2354,22 @@ class AsyncIndex(_BaseIndex):
             >>>     index = client.index("movies")
             >>>     await index.update_documents_in_batches(documents)
         """
+        if concurrency_limit:
+            async with asyncio.Semaphore(concurrency_limit):
+                if not use_task_groups():
+                    batches = [
+                        self.update_documents(x, primary_key, compress=compress)
+                        for x in _batch(documents, batch_size)
+                    ]
+                    return await asyncio.gather(*batches)
+
+                async with asyncio.TaskGroup() as tg:  # type: ignore[attr-defined]
+                    tasks = [
+                        tg.create_task(self.update_documents(x, primary_key, compress=compress))
+                        for x in _batch(documents, batch_size)
+                    ]
+                return [x.result() for x in tasks]
+
         if not use_task_groups():
             batches = [
                 self.update_documents(x, primary_key, compress=compress)
@@ -2402,6 +2506,7 @@ class AsyncIndex(_BaseIndex):
         csv_delimiter: str | None = None,
         combine_documents: bool = True,
         compress: bool = False,
+        concurrency_limit: int | None = None,
     ) -> list[TaskInfo]:
         """Load all json files from a directory and update the documents.
 
@@ -2420,6 +2525,9 @@ class AsyncIndex(_BaseIndex):
             combine_documents: If set to True this will combine the documents from all the files
                 before indexing them. Defaults to True.
             compress: If set to True the data will be sent in gzip format. Defaults to False.
+            concurrency_limit: If set this will limit the number of batches that will be sent
+                concurrently. This can be helpful if you find you are overloading the Meilisearch
+                server with requests. Defaults to None.
 
         Returns:
 
@@ -2462,6 +2570,7 @@ class AsyncIndex(_BaseIndex):
                 batch_size=batch_size,
                 primary_key=primary_key,
                 compress=compress,
+                concurrency_limit=concurrency_limit,
             )
 
         if not use_task_groups():
@@ -2479,6 +2588,7 @@ class AsyncIndex(_BaseIndex):
                             batch_size=batch_size,
                             primary_key=primary_key,
                             compress=compress,
+                            concurrency_limit=concurrency_limit,
                         )
                     )
 
@@ -2509,6 +2619,7 @@ class AsyncIndex(_BaseIndex):
                             batch_size=batch_size,
                             primary_key=primary_key,
                             compress=compress,
+                            concurrency_limit=concurrency_limit,
                         )
                     else:
                         tasks.append(
@@ -2518,6 +2629,7 @@ class AsyncIndex(_BaseIndex):
                                     batch_size=batch_size,
                                     primary_key=primary_key,
                                     compress=compress,
+                                    concurrency_limit=concurrency_limit,
                                 )
                             )
                         )
@@ -2576,6 +2688,7 @@ class AsyncIndex(_BaseIndex):
         batch_size: int = 1000,
         primary_key: str | None = None,
         compress: bool = False,
+        concurrency_limit: int | None = None,
     ) -> list[TaskInfo]:
         """Updates documents form a json file in batches to reduce RAM usage with indexing.
 
@@ -2587,6 +2700,9 @@ class AsyncIndex(_BaseIndex):
             primary_key: The primary key of the documents. This will be ignored if already set.
                 Defaults to None.
             compress: If set to True the data will be sent in gzip format. Defaults to False.
+            concurrency_limit: If set this will limit the number of batches that will be sent
+                concurrently. This can be helpful if you find you are overloading the Meilisearch
+                server with requests. Defaults to None.
 
         Returns:
 
@@ -2615,6 +2731,7 @@ class AsyncIndex(_BaseIndex):
             batch_size=batch_size,
             primary_key=primary_key,
             compress=compress,
+            concurrency_limit=concurrency_limit,
         )
 
     async def update_documents_from_raw_file(
@@ -2692,7 +2809,7 @@ class AsyncIndex(_BaseIndex):
         else:
             url = self._documents_url
 
-        async with aiofiles.open(upload_path, "r") as f:
+        async with aiofiles.open(upload_path) as f:
             data = await f.read()
 
         response = await self._http_requests.put(
@@ -2935,13 +3052,16 @@ class AsyncIndex(_BaseIndex):
         return result
 
     async def delete_documents_in_batches_by_filter(
-        self, filters: list[str | list[str | list[str]]]
+        self, filters: list[str | list[str | list[str]]], concurrency_limit: int | None = None
     ) -> list[TaskInfo]:
         """Delete batches of documents from the index by filter.
 
         Args:
 
             filters: A list of filter value information.
+            concurrency_limit: If set this will limit the number of batches that will be sent
+                concurrently. This can be helpful if you find you are overloading the Meilisearch
+                server with requests. Defaults to None.
 
         Returns:
 
@@ -2964,6 +3084,20 @@ class AsyncIndex(_BaseIndex):
             >>>         ]
             >>>     )
         """
+        if concurrency_limit:
+            async with asyncio.Semaphore(concurrency_limit):
+                if not use_task_groups():
+                    tasks = [self.delete_documents_by_filter(filter) for filter in filters]
+                    return await asyncio.gather(*tasks)
+
+                async with asyncio.TaskGroup() as tg:  # type: ignore[attr-defined]
+                    tg_tasks = [
+                        tg.create_task(self.delete_documents_by_filter(filter))
+                        for filter in filters
+                    ]
+
+                return [x.result() for x in tg_tasks]
+
         if not use_task_groups():
             tasks = [self.delete_documents_by_filter(filter) for filter in filters]
             return await asyncio.gather(*tasks)
@@ -8426,7 +8560,7 @@ async def _async_load_documents_from_file(
         with open(file_path) as f:  # noqa: ASYNC101 ASYNC230
             return [await loop.run_in_executor(None, partial(json_handler.loads, x)) for x in f]
 
-    async with aiofiles.open(file_path, mode="r") as f:  # type: ignore
+    async with aiofiles.open(file_path) as f:  # type: ignore
         data = await f.read()  # type: ignore
         documents = await loop.run_in_executor(None, partial(json_handler.loads, data))
 

{meilisearch_python_sdk-3.3.0.dist-info → meilisearch_python_sdk-3.5.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: meilisearch-python-sdk
-Version: 3.3.0
+Version: 3.5.0
 Summary: A Python client providing both async and sync support for the Meilisearch API
 Project-URL: repository, https://github.com/sanders41/meilisearch-python-sdk
 Project-URL: homepage, https://github.com/sanders41/meilisearch-python-sdk
@@ -42,7 +42,7 @@ Requires-Python: >=3.9
 Requires-Dist: aiofiles>=0.7
 Requires-Dist: camel-converter>=1.0.0
 Requires-Dist: eval-type-backport>=0.2.0; python_version < '3.10'
-Requires-Dist: httpx>=0.17
+Requires-Dist: httpx[http2]>=0.17
 Requires-Dist: pydantic>=2.0.0
 Requires-Dist: pyjwt>=2.3.0
 Provides-Extra: all

{meilisearch_python_sdk-3.3.0.dist-info → meilisearch_python_sdk-3.5.0.dist-info}/RECORD

@@ -1,12 +1,12 @@
 meilisearch_python_sdk/__init__.py,sha256=SB0Jlm6FwT13J9xasZKseZzTWBk0hkfe1CWyWmIIZnE,258
-meilisearch_python_sdk/_client.py,sha256=HLbU0GLAh5TB-L9GuFHrCQEipiQusa94hyqak09ULOU,69865
+meilisearch_python_sdk/_client.py,sha256=8QSMTzpyIZQ9dOIsvhPNtvdbBfIk8e1LCBtE5V9Qdak,70166
 meilisearch_python_sdk/_http_requests.py,sha256=TwpqsOvfgaJ1lQXwam1q1_UC6NvRWy4m9W3c5KNe0RI,6741
 meilisearch_python_sdk/_task.py,sha256=dB0cpX1u7HDM1OW_TC8gSiGJe985bNCz7hPMZW_qogY,12352
 meilisearch_python_sdk/_utils.py,sha256=k6SYMJSiVjfF-vlhQRMaE1ziJsVf5FrL94mFwrMfdLY,957
-meilisearch_python_sdk/_version.py,sha256=PzSl7vO5fA6PZhs4iV8xBCu7yMgUgHHWOABscpEi88A,18
-meilisearch_python_sdk/decorators.py,sha256=KpS5gAgks28BtPMZJumRaXfgXK4A3QNVPR8Z4BpZC0g,8346
+meilisearch_python_sdk/_version.py,sha256=xrNY8UGfJPo6Zzc_jnHNnSunmOWjJ2loI3iSY-UCmY8,18
+meilisearch_python_sdk/decorators.py,sha256=hNrMvuLJKPNQDULkL1yMZYG7A9OVYbT7nass4URtEZM,8684
 meilisearch_python_sdk/errors.py,sha256=0sAKYt47-zFpKsEU6W8Qnvf4uHBynKtlGPpPl-5laSA,2085
-meilisearch_python_sdk/index.py,sha256=2y2_gDuxtflmnduPtm7LdTZzrn-sOvFU0rAZDFmcP6c,326043
+meilisearch_python_sdk/index.py,sha256=cr1VpUih5LabDL0yTiBP7wJnDIOkZd21ewDTojcdah0,333340
 meilisearch_python_sdk/json_handler.py,sha256=q_87zSnJfDNuVEI9cEvuOQOGBC7AGWJMEqCh2kGAAqA,2107
 meilisearch_python_sdk/plugins.py,sha256=YySzTuVr4IrogTgrP8q-gZPsew8TwedopjWnTj5eV48,3607
 meilisearch_python_sdk/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -20,7 +20,7 @@ meilisearch_python_sdk/models/search.py,sha256=LH64_TunWxfCJOhxMCnFA-bMZOf7fVx6s
 meilisearch_python_sdk/models/settings.py,sha256=A8SocaQldrdo1chvxhS522zZR4foJcvZy7Cg2GiBi_M,3968
 meilisearch_python_sdk/models/task.py,sha256=P3NLaZhrY8H02Q9lDEkoq-3Z6_qGESglOxs4dNRyMWg,2100
 meilisearch_python_sdk/models/version.py,sha256=YDu-aj5H-d6nSaWRTXzlwWghmZAoiknaw250UyEd48I,215
-meilisearch_python_sdk-3.3.0.dist-info/METADATA,sha256=KRcd87AO1_lX6bCTn511e6XQSpoSncX2bTbkCZXIpC0,9703
-meilisearch_python_sdk-3.3.0.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
-meilisearch_python_sdk-3.3.0.dist-info/licenses/LICENSE,sha256=xVzevI1TrlKfM0plmJ7vfK1Muu0V9n-dGE8RnDrOFlM,1069
-meilisearch_python_sdk-3.3.0.dist-info/RECORD,,
+meilisearch_python_sdk-3.5.0.dist-info/METADATA,sha256=nyBheqJZbF0Z5kbO0Pber_hSqMk3rgPbuk8kq7ifAIM,9710
+meilisearch_python_sdk-3.5.0.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
+meilisearch_python_sdk-3.5.0.dist-info/licenses/LICENSE,sha256=xVzevI1TrlKfM0plmJ7vfK1Muu0V9n-dGE8RnDrOFlM,1069
+meilisearch_python_sdk-3.5.0.dist-info/RECORD,,