rudi-node-write 0.1.0__tar.gz → 1.0.0__tar.gz

{rudi-node-write-0.1.0/src/rudi_node_write.egg-info → rudi_node_write-1.0.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: rudi-node-write
-Version: 0.1.0
+Version: 1.0.0
 Summary: Use the internal API of a RUDI Producer node
 Author-email: Olivier Martineau <olivier.martineau@irisa.fr>
 Maintainer-email: Olivier Martineau <olivier.martineau@irisa.fr>
@@ -9,49 +9,48 @@ Project-URL: Homepage, https://github.com/OlivierMartineau/rudi-node-write
 Project-URL: Documentation, https://app.swaggerhub.com/apis/OlivierMartineau/RudiProducer-InternalAPI
 Project-URL: Changelog, https://github.com/OlivierMartineau/rudi-node-write/blob/release/CHANGELOG.md
 Project-URL: Repository, https://github.com/OlivierMartineau/rudi-node-write
-Keywords: rudi-node-write,rudi-node-get,RUDI,producer node,RUDI node,open-data,Univ. Rennes
-Classifier: Programming Language :: Python :: 3.10
+Keywords: rudi-node-write,rudi-node-put,RUDI,producer node,RUDI node,open-data,Univ. Rennes
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Natural Language :: English
 Classifier: Operating System :: OS Independent
 Classifier: License :: OSI Approved :: European Union Public Licence 1.2 (EUPL 1.2)
-Requires-Python: >=3.10
+Requires-Python: >=3.11
 Description-Content-Type: text/markdown
 License-File: LICENCE.md
+Requires-Dist: beautifulsoup4==4.12.3
+Requires-Dist: chardet==5.2.0
+Requires-Dist: deepdiff==7.0.1
+Requires-Dist: defusedxml==0.7.1
+Requires-Dist: puremagic==1.24
 
 [![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
-
-Caution: this librairy is still a work in progress.
+[![Checked with pyright](https://microsoft.github.io/pyright/img/pyright_badge.svg)](https://microsoft.github.io/pyright/)
 
 # RUDI Node tools: _rudi-node-write_ library
 
 This library offers tools to take advantage of
 the [internal API](https://app.swaggerhub.com/apis/OlivierMartineau/RudiProducer-InternalAPI) of a RUDI Producer node (
 also
-referred as RUDI node).
+referred as RUDI node), through the API of the backend of the user interface, the "Producer node manager" or "Prodmanager" module.
 
 ## Installation
 
 ```bash
-$ pip install rudi_node_write
+python3 -m venv .venv
+source .venv/bin/activate
+pip install rudi_node_write
 ```
 
-## Usage
-
-```python
-from rudi_node_write.connectors.io_rudi_api_write import RudiNodeApiConnector
+## Usage: RudiNodeWriter
 
-rudi_api = RudiNodeApiConnector('https://bacasable.fenix.rudi-univ-rennes1.fr')
-print(rudi_api.metadata_count)
-print(len(rudi_api.metadata_list))
-print(rudi_api.producer_names)
-print(rudi_api.find_metadata_with_media_name('toucan.jpg'))
-
-```
+The Jupyter notebook [`README.ipynb`](https://github.com/OlivierMartineau/rudi-node-write/blob/release/README.ipynb) details how to use the library through the [`RudiNodeWriter`](https://github.com/OlivierMartineau/rudi-node-write/blob/release/src/rudi_node_write/rudi_node_writer.py) object.
 
 ## Testing
 
+The [tests](https://github.com/OlivierMartineau/rudi-node-write/tree/release/tests) can be analyzed for further
+information about how to call the API
+
 ```bash
 $ pytest
 ```

rudi_node_write-1.0.0/README.md

@@ -0,0 +1,30 @@
+[![Code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
+[![Checked with pyright](https://microsoft.github.io/pyright/img/pyright_badge.svg)](https://microsoft.github.io/pyright/)
+
+# RUDI Node tools: _rudi-node-write_ library
+
+This library offers tools to take advantage of
+the [internal API](https://app.swaggerhub.com/apis/OlivierMartineau/RudiProducer-InternalAPI) of a RUDI Producer node (
+also
+referred as RUDI node), through the API of the backend of the user interface, the "Producer node manager" or "Prodmanager" module.
+
+## Installation
+
+```bash
+python3 -m venv .venv
+source .venv/bin/activate
+pip install rudi_node_write
+```
+
+## Usage: RudiNodeWriter
+
+The Jupyter notebook [`README.ipynb`](https://github.com/OlivierMartineau/rudi-node-write/blob/release/README.ipynb) details how to use the library through the [`RudiNodeWriter`](https://github.com/OlivierMartineau/rudi-node-write/blob/release/src/rudi_node_write/rudi_node_writer.py) object.
+
+## Testing
+
+The [tests](https://github.com/OlivierMartineau/rudi-node-write/tree/release/tests) can be analyzed for further
+information about how to call the API
+
+```bash
+$ pytest
+```

{rudi-node-write-0.1.0 → rudi_node_write-1.0.0}/pyproject.toml

@@ -4,21 +4,22 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "rudi-node-write"
-version = "0.1.0"
+version = "1.0.0"
 authors = [{ name = "Olivier Martineau", email = "olivier.martineau@irisa.fr" }]
 maintainers = [{ name = "Olivier Martineau", email = "olivier.martineau@irisa.fr" }]
 description = "Use the internal API of a RUDI Producer node"
 readme = "README.md"
-requires-python = ">=3.10"
+requires-python = ">=3.11"
 license = { text = "EUPL-1.2" }
-classifiers = ["Programming Language :: Python :: 3.10",
-    "Programming Language :: Python :: 3.11",
+classifiers = ["Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
     "Natural Language :: English",
     "Operating System :: OS Independent",
     "License :: OSI Approved :: European Union Public Licence 1.2 (EUPL 1.2)",
 ]
-keywords = ["rudi-node-write", "rudi-node-get", "RUDI", "producer node", "RUDI node", "open-data", "Univ. Rennes"]
+keywords = ["rudi-node-write", "rudi-node-put", "RUDI", "producer node", "RUDI node", "open-data", "Univ. Rennes"]
+dynamic = ["dependencies"]
+# dependencies = ["beautifulsoup4", "chardet", "deepdiff", "defusedxml", "puremagic"]
 
 [project.urls]
 Homepage = "https://github.com/OlivierMartineau/rudi-node-write"
@@ -34,7 +35,7 @@ target-version = ['py311']
 # ----- Tool: commitizen
 [tool.commitizen]
 name = "cz_conventional_commits"
-version = "0.1.0"
+version = "1.0.0"
 version_files = ["pyproject.toml:version"]
 
 # ----- Tool: pytest
@@ -42,4 +43,24 @@ version_files = ["pyproject.toml:version"]
 pythonpath = ["src"]
 norecursedirs = ["*.egg", ".eggs", "dist", "build"]
 filterwarnings = ["ignore:.*pkg_resources.*:DeprecationWarning"]
-addopts = ["--cov=rudi_node_write", "--cov-report=term-missing", "--cov-report=html"]
+addopts = ["--cov=rudi_node_write", "--cov-report=term-missing", "--cov-report=html:reports/html_dir", "--cov-report=xml:reports/coverage.xml" ]
+
+# ----- Tool: setuptools
+[tool.setuptools.packages.find]
+where = ["src"]
+include = ["rudi_node_write*"]
+
+[tool.setuptools.dynamic]
+dependencies = {file = ["requirements.txt"]}
+
+# ----- Tool: pyright
+[tool.pyright]
+exclude = ["**/node_modules", "**/__pycache__"]
+include = ["src"]
+pythonVersion = "3.11"
+reportMissingImports = "true"
+reportDuplicateImport = "true"
+reportUnusedImport = "true"
+reportImportCycles = "true"
+venvPath = "."
+venv = ".venv"

rudi_node_write-1.0.0/requirements.txt

@@ -0,0 +1,7 @@
+beautifulsoup4==4.12.3
+chardet==5.2.0
+deepdiff==7.0.1
+defusedxml==0.7.1
+puremagic==1.24
+# ordered-set==4.1.0 # Installed as dependency for deepdiff
+# soupsieve==2.5 # Installed as dependency for beautifulsoup4

{rudi-node-write-0.1.0 → rudi_node_write-1.0.0}/src/rudi_node_write/conf/meta_defaults.py

@@ -13,7 +13,7 @@ ODS_API_VERSION: str = "2.1"
 
 # Default contact when none is provided in the source metadata
 # Structure: https://app.swaggerhub.com/apis/OlivierMartineau/RUDI-PRODUCER/1.3.0#/Contact
-DEFAULT_CONTACT: dict = dict(contact_name="Rudi node admin", email="community@rudi-univ-rennes1.fr")
+DEFAULT_CONTACT: dict = {"contact_name": "Rudi node admin", "email": "community@rudi-univ-rennes1.fr"}
 
 # Default producer when none is provided in the source metadata
 # Structure: https://app.swaggerhub.com/apis/OlivierMartineau/RUDI-PRODUCER/1.3.0#/Organization

{rudi-node-write-0.1.0 → rudi_node_write-1.0.0}/src/rudi_node_write/connectors/io_connector.py

@@ -4,18 +4,17 @@ from typing import Literal, get_args, BinaryIO, TextIO
 from urllib.parse import urlsplit
 
 from rudi_node_write.rudi_types.rudi_const import check_is_literal
-from rudi_node_write.utils.dict_utils import is_dict
+from rudi_node_write.rudi_types.serializable import Serializable
+from rudi_node_write.utils.dict_utils import is_dict, safe_get_key
 from rudi_node_write.utils.err import HttpError
 from rudi_node_write.utils.log import log_d_if, log_e, log_d
 from rudi_node_write.utils.str_utils import slash_join
-from rudi_node_write.utils.url_utils import url_encode_req_params
+from rudi_node_write.utils.url_utils import get_response_cookies, url_encode_req_params
 
-HttpRequestMethod = Literal["GET", "PUT", "DEL", "POST"]
+HttpRequestMethod = Literal["GET", "PUT", "DELETE", "POST"]
 HTTP_REQUEST_METHODS = get_args(HttpRequestMethod)
 
-CONTENT_TYPE_KEY = "Content-Type"
-
-DEFAULT_HEADERS = {CONTENT_TYPE_KEY: "text/plain; ", "Accept": "application/json"}
+DEFAULT_HEADERS = {"Content-Type": "text/plain; ", "Accept": "application/json"}
 
 STATUS = "status"
 REDIRECTION = "redirection"
@@ -26,9 +25,9 @@ def https_download(resource_url: str, headers=None, should_show_debug_line: bool
         headers = DEFAULT_HEADERS
     here = "https_download"
     (scheme, netloc, path, query, fragment) = urlsplit(resource_url)
-    if scheme != "https":
+    if scheme not in ("https", "http"):
         raise NotImplementedError(f"only HTTPS protocol is supported, cannot treat this url: {resource_url}")
-    connection = HTTPSConnection(netloc)
+    connection = (HTTPSConnection if scheme == "https" else HTTPConnection)(netloc)
 
     connection.request(method="GET", url=resource_url, headers=headers)
     response = connection.getresponse()
@@ -53,18 +52,18 @@ class Connector:
         self.connection = None
 
         self._set_url(server_url)
+        self._cookies = None
 
     def _set_url(self, server_url: str):
-        fun = f"super.{self.class_name}._set_url"
+        here = f"super.{self.class_name}._set_url"
         (scheme, netloc, path, query, fragment) = urlsplit(server_url)
-        if scheme != "http" and scheme != "https":
+        if scheme not in ("https", "http"):
             raise NotImplementedError(f"only http and https are supported, got '{scheme}'")
-        # self.scheme = scheme
-        self.scheme = "https"
+        self.scheme = scheme
         self.host = netloc
         self.path = path
         self.base_url = slash_join(f"{self.scheme}://{self.host}", self.path)
-        log_d(fun, "base_url", self.base_url)
+        log_d(here, "base_url", self.base_url)
 
     @property
     def class_name(self):
@@ -76,17 +75,23 @@ class Connector:
     def full_path(self, relative_url: str = "/"):
         return slash_join("/", self.path, url_encode_req_params(relative_url))
 
-    def test_connection(self):
-        return self.request()
+    def test_connection(self) -> bool | str | dict:
+        return self.request()  # type: ignore
 
     def close_connection(self):
+        if self.connection is None:
+            raise ConnectionAbortedError("The connection should be alive still")
         try:
             self.connection.close()
         except Exception as e:  # pragma: no cover
             log_e(self.class_name, "close_connection ERROR", e)
 
     def download(
-        self, relative_url: str, headers: dict = None, keep_alive: bool = False, should_log_response: bool = False
+        self,
+        relative_url: str,
+        headers: dict | None = None,
+        keep_alive: bool = False,
+        should_log_response: bool = False,
     ):
         """
         Download a file on the connector server
@@ -98,6 +103,8 @@ class Connector:
         :return: a status
         """
         here = f"{self.class_name}.download"
+        if self.host is None:
+            raise AttributeError("The host was not defined")
         if headers is None:
             headers = DEFAULT_HEADERS
 
@@ -122,11 +129,12 @@ class Connector:
         self,
         relative_url: str = "/",
         req_method: HttpRequestMethod = "GET",
-        body: dict | str | BinaryIO | TextIO = None,
+        body: dict | str | list | Serializable | BinaryIO | TextIO | None = None,
         headers=None,
         keep_alive: bool = False,
+        should_log_request: bool = False,
         should_log_response: bool = False,
-    ) -> (str, dict):
+    ):
         """
         Send a http(s) request
         :param relative_url: a relative URL that will be joined to the connector's base URL to form the request URL
@@ -138,30 +146,35 @@ class Connector:
         :param should_log_response: True if some log lines should be displayed (defaults to False).
         :return: the data returned from the request
         """
-        fun = f"{self.class_name}.request"
+        here = f"{self.class_name}.request"
 
         check_is_literal(req_method, HTTP_REQUEST_METHODS, "incorrect type for request method")
 
         if headers is None:
             headers = DEFAULT_HEADERS
-        if is_dict(body):
-            headers[CONTENT_TYPE_KEY] = "application/json"
-            body = dumps(body)
+        if isinstance(body, dict):
+            headers["Content-Type"] = "application/json"
+            body_str = dumps(body)
+        elif isinstance(body, Serializable):
+            headers["Content-Type"] = "application/json"
+            body_str = body.to_json_str(ensure_ascii=True)
+        elif isinstance(body, list):
+            body_str = str(list)
+        else:
+            body_str = body
 
         path_url = self.full_path(relative_url)
+        if self.host is None:
+            raise ConnectionError("The connector host should be defined")
+        self.connection = HTTPConnection(self.host) if self.scheme == "http" else HTTPSConnection(self.host)
 
-        if self.scheme == "http":  # pragma: no cover
-            self.connection = HTTPConnection(self.host)
-        else:
-            self.connection = HTTPSConnection(self.host)
-        log_d(fun, req_method, self.full_url(relative_url))
+        log_d_if(should_log_request, here, req_method, self.full_url(relative_url))
 
         try:
-            # log_d(fun, "request details", "method=", req_method, "url=", path_url, "body=", body, "headers=", headers)
-            self.connection.request(method=req_method, url=path_url, body=body, headers=headers)
+            self.connection.request(method=req_method, url=path_url, body=body_str, headers=headers)  # type: ignore
         except ConnectionRefusedError as e:
-            log_e(fun, "Error on request", req_method, self.full_url(relative_url))
-            log_e(fun, "ERR", e)
+            log_e(here, "Error on request", req_method, self.full_url(relative_url))
+            log_e(here, "ERR", e)
             raise e
         return self.parse_response(
             relative_url=relative_url,
@@ -178,9 +191,12 @@ class Connector:
         should_log_response: bool = True,
     ):
         """Basic parsing of the result"""
-        fun = f"{self.class_name}.parse_response"
+        here = f"{self.class_name}.parse_response"
+        if self.connection is None:
+            raise ConnectionError("The connector HTTPSConnection should be defined")
         response = self.connection.getresponse()
-        # log_d(fun, "Response", response.getcode(), response.getheaders(), response.info())
+        self._cookies = get_response_cookies(response)
+
         if response.status in [301, 302]:
             return {STATUS: response.status, REDIRECTION: response.getheader("location")}
         if (
@@ -191,25 +207,26 @@ class Connector:
             return None
 
         rdata = response.read()
-        # log_d(fun, "rdata", rdata)
+        # log_d(here, "rdata", rdata)
         try:
             response_data = loads(rdata)
-            log_d_if(should_log_response, fun, "Response is a JSON", response_data)
+            log_d_if(should_log_response, here, "Response is a JSON", response_data)
         except (TypeError, JSONDecodeError):
             response_data = repr(rdata)
-            log_d_if(should_log_response, fun, "Response is not a JSON", response_data)
+            log_d_if(should_log_response, here, "Response is not a JSON", response_data)
         if not keep_alive:
             self.close_connection()
 
         if type(response_data) is str:
-            log_d_if(should_log_response, fun, "Response is a string", response_data)
+            log_d_if(should_log_response, here, "Response is a string", response_data)
             if response.status == 200:
                 return rdata.decode("utf8")
         if response.status == 200:
             return response_data
         if response.status >= 400:
-            log_e(fun, "Connection error", response_data)
-            log_e(fun, "Request in error", req_method, self.full_url(relative_url))
+            log_e(here, "Connection error", response_data)
+            log_e(here, "Request in error", req_method, self.full_url(relative_url))
+            # log_e(here, "Headers", response.headers)
             raise HttpError(response_data, req_method, self.base_url, relative_url)
 
 
@@ -217,13 +234,31 @@ if __name__ == "__main__":  # pragma: no cover
     tests = "Connector tests"
     connector = Connector("https://bacasable.fenix.rudi-univ-rennes1.fr/api/version")
     log_d(tests, "testing connection, got version", connector.test_connection())
-    # data = connector.request(relative_url="resources?limit=1")
-    # print(data)
 
     connector = Connector("https://bacasable.fenix.rudi-univ-rennes1.fr/api/v1")
-    res = connector.request("resources")
-
-    log_d(tests, "number of resources declared", res["total"])
+    metadata_list = connector.request("resources")
+    if isinstance(metadata_list, dict) and metadata_list["total"] is not None:
+        log_d(tests, "number of resources declared", f"{metadata_list['total']}")
+    else:
+        log_e(
+            tests,
+            f"An error occurred, result should be a dict with 'total' and 'items' properties, got {metadata_list}",
+        )
 
-    url = "https://bacasable.fenix.rudi-univ-rennes1.fr/media/download/b086c7b2-bd6d-401f-86f5-f1f207023bae"
-    log_d("https_utils", url, https_download(url))
+    url = "resources?available_formats.file_storage_status=available&fields=available_formats"
+    meta_list: list = connector.request(url)["items"]  # type: ignore
+    # print(tests, "media_list", meta_list)
+    available_files = []
+    for meta in meta_list:
+        for media in meta["available_formats"]:
+            # log_d(tests, "media", media)
+            file_storage_status = safe_get_key(media, "file_storage_status")
+            if file_storage_status == "available":
+                available_files.append(media)
+
+    if len(available_files) == 0:
+        print("no media available, quiting")
+    else:
+        media_1 = available_files[0]
+        url = media_1["connector"]["url"]
+        log_d("https_utils", url, https_download(url))