v0.5.0 (#153)

Changes proposed in this pull request:

 * Adjusted AI example to use automatic model download feature
 * Adjusted Makefiles of all examples
 * Fixed missing default icon for `files_dropdown_menu` API
 * Updated README.md

---------

Signed-off-by: Alexander Piskun <[email protected]>

Author: bigcat88

.run/Skeleton.run.xml renamed to .run/Skeleton (28).run.xml

@@ -1,15 +1,15 @@
 <component name="ProjectRunConfigurationManager">
-  <configuration default="false" name="Skeleton" type="PythonConfigurationType" factoryName="Python">
+  <configuration default="false" name="Skeleton (28)" type="PythonConfigurationType" factoryName="Python">
     <module name="nc_py_api" />
     <option name="INTERPRETER_OPTIONS" value="" />
     <option name="PARENT_ENVS" value="true" />
     <envs>
+      <env name="PYTHONUNBUFFERED" value="1" />
       <env name="APP_ID" value="skeleton" />
       <env name="APP_PORT" value="9030" />
       <env name="APP_SECRET" value="12345" />
       <env name="APP_VERSION" value="1.0.0" />
       <env name="NEXTCLOUD_URL" value="http://nextcloud.local/index.php" />
-      <env name="PYTHONUNBUFFERED" value="1" />
     </envs>
     <option name="SDK_HOME" value="" />
     <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />

.run/TalkBot.run.xml renamed to .run/TalkBot (28).run.xml

@@ -1,15 +1,15 @@
 <component name="ProjectRunConfigurationManager">
-  <configuration default="false" name="TalkBot" type="PythonConfigurationType" factoryName="Python">
+  <configuration default="false" name="TalkBot (28)" type="PythonConfigurationType" factoryName="Python">
     <module name="nc_py_api" />
     <option name="INTERPRETER_OPTIONS" value="" />
     <option name="PARENT_ENVS" value="true" />
     <envs>
+      <env name="PYTHONUNBUFFERED" value="1" />
       <env name="APP_ID" value="talk_bot" />
       <env name="APP_PORT" value="9032" />
       <env name="APP_SECRET" value="12345" />
       <env name="APP_VERSION" value="1.0.0" />
       <env name="NEXTCLOUD_URL" value="http://nextcloud.local/index.php" />
-      <env name="PYTHONUNBUFFERED" value="1" />
     </envs>
     <option name="SDK_HOME" value="" />
     <option name="WORKING_DIRECTORY" value="$PROJECT_DIR$" />

.run/TalkBotMulti.run.xml

@@ -2,15 +2,16 @@
 
 All notable changes to this project will be documented in this file.
 
-## [0.4.1 - 2023-10-17]
+## [0.5.0 - 2023-10-23]
 
 ### Added
 
-- Support for the new AppAPI endpoint `/init` and automatically downloading models from `huggingface`. #151
+- Support for the new `/init` AppAPI endpoint and the ability to automatically load models from `huggingface`. #151
 
 ### Changed
 
 - All examples were adjusted to changes in AppAPI.
+- The examples now use FastAPIs `lifespan` instead of the deprecated `on_event`.
 
 ## [0.4.0 - 2023-10-15]
 

CHANGELOG.md

@@ -57,6 +57,38 @@ as long as it doesn't involve calls that require user password verification.
 
 **NextcloudApp** avalaible only from Nextcloud 27.1.2 and greater version with installed **AppAPI**.
 
+### Nextcloud skeleton app in Python
+
+```python3
+from contextlib import asynccontextmanager
+
+from fastapi import FastAPI
+
+from nc_py_api import NextcloudApp
+from nc_py_api.ex_app import LogLvl, run_app, set_handlers
+
+
+@asynccontextmanager
+async def lifespan(_app: FastAPI):
+    set_handlers(APP, enabled_handler)
+    yield
+
+
+APP = FastAPI(lifespan=lifespan)
+
+
+def enabled_handler(enabled: bool, nc: NextcloudApp) -> str:
+    if enabled:
+        nc.log(LogLvl.WARNING, "Hello from nc_py_api.")
+    else:
+        nc.log(LogLvl.WARNING, "Bye bye from nc_py_api.")
+    return ""
+
+
+if __name__ == "__main__":
+    run_app("main:APP", log_level="trace")
+```
+
 ### Support
 
 You can support us in several ways:
@@ -81,18 +113,3 @@ You can support us in several ways:
   - [Issues](https://github.com/cloud-py-api/nc_py_api/issues)
   - [Setting up dev environment](https://cloud-py-api.github.io/nc_py_api/DevSetup.html)
 - [Changelog](https://github.com/cloud-py-api/nc_py_api/blob/main/CHANGELOG.md)
-
-### Motivation
-
-_Python's language, elegant and clear,_<br>
-_Weaves logic's threads without fear,_<br>
-_And in the sky, where clouds take form,_<br>
-_Nextcloud emerges, a digital norm._<br>
-
-_Together they stand, a duo bright,_<br>
-_Python and Nextcloud, day and night,_<br>
-_In a digital dance, they guide and sail,_<br>
-_Shaping tomorrow, where new ideas prevail._<br>
-
-#### **Know that we are always here to support and assist you on your journey.**
-### P.S: **_Good luck, and we hope you have fun!_**

README.md

@@ -12,12 +12,12 @@ Skeleton
 
 What's going on in the skeleton?
 
-First, it's important to understand that an external application acts more like a microservice, with its endpoints being called by Nextcloud.
+In `FastAPI lifespan <https://fastapi.tiangolo.com/advanced/events/?h=lifespan#lifespan>`_ we call the ``set_handlers`` function to further process the application installation logic.
 
-Therefore, when the application receives a request at the endpoint ``/enable``,
-it should register all its functionalities in the cloud and wait for requests from Nextcloud.
+Since this is a simple skeleton application, we only define the ``/enable`` endpoint.
 
-.. note:: This doesn't apply to system applications, which will be covered in the next chapter.
+When the application receives a request at the endpoint ``/enable``,
+it should register all its functionalities in the cloud and wait for requests from Nextcloud.
 
 So, calling:
 
@@ -54,7 +54,7 @@ Here they are:
 
 * APP_ID - ID of the application.
 * APP_PORT - Port on which application listen for the requests from the Nextcloud.
-* APP_SECRET - Secret for ``hmac`` signature generation.
+* APP_SECRET - Shared secret between Nextcloud and Application.
 * APP_VERSION - Version of the application.
 * AA_VERSION - Version of the AppAPI.
 * NEXTCLOUD_URL - URL at which the application can access the Nextcloud API.
@@ -67,7 +67,7 @@ After launching your application, execute the following command in the Nextcloud
 
     php occ app_api:app:register YOUR_APP_ID manual_install --json-info \
         "{\"appid\":\"YOUR_APP_ID\",\"name\":\"YOUR_APP_DISPLAY_NAME\",\"daemon_config_name\":\"manual_install\",\"version\":\"YOU_APP_VERSION\",\"secret\":\"YOUR_APP_SECRET\",\"host\":\"host.docker.internal\",\"scopes\":{\"required\":[2, 10, 11],\"optional\":[30, 31, 32, 33]},\"port\":SELECTED_PORT,\"protocol\":\"http\",\"system_app\":0}" \
-        -e --force-scopes
+        --force-scopes
 
 You can see how **nc_py_api** registers in ``scripts/dev_register.sh``.
 
@@ -79,8 +79,7 @@ Examples for such Makefiles can be found in this repository:
 `ToGif <https://github.com/cloud-py-api/nc_py_api/blob/main/examples/as_app/to_gif/Makefile>`_ ,
 `nc_py_api <https://github.com/cloud-py-api/nc_py_api/blob/main/scripts/dev_register.sh>`_
 
-During the execution of `php occ app_api:app:register`, the **enabled_handler** will be called,
-as we pass the flag ``-e``, meaning ``enable after registration``.
+During the execution of `php occ app_api:app:register`, the **enabled_handler** will be called
 
 This is likely all you need to start debugging and developing an application for Nextcloud.
 

docs/NextcloudApp.rst

@@ -37,7 +37,7 @@ Requirements
 We opt for the latest version of the Transformers library.
 Because the example was developed on a Mac, we ended up using Torchvision.
 
-`If you're working solely with Nvidia, you're free to use TensorFlow instead of PyTorch.`
+`You're free to use TensorFlow instead of PyTorch.`
 
 Next, we integrate the latest version of `nc_py_api` to minimize code redundancy and focus on the application's logic.
 
@@ -52,63 +52,20 @@ We specify the model name globally so that we can easily change the model name i
 
 **When Should We Download the Language Model?**
 
-Although the example uses the smallest model available, weighing in at 300 megabytes, it's common knowledge that larger language models can be substantially bigger.
-Downloading such models should not begin when a processing request is already received.
+To make process of initializing applications more robust, separate logic was introduced, with an ``/init`` endpoint.
 
-So we have two options:
-
-* Heartbeat
-* enabled_handler
-
-This can't be accomplished in the **app on/off handler** as Nextcloud expects an immediate response regarding the app's operability.
-
-Thus, we place the model downloading within the Heartbeat:
-
-.. code-block::
-
-    # Thread that performs model download.
-    def download_models():
-        pipeline("text2text-generation", model=MODEL_NAME)  # this will download model
-
-
-    def heartbeat_handler() -> str:
-        global MODEL_INIT_THREAD
-        print("heartbeat_handler: called")  # for debug
-        # if it is the first heartbeat, then start background thread to download a model
-        if MODEL_INIT_THREAD is None:
-            MODEL_INIT_THREAD = Thread(target=download_models)
-            MODEL_INIT_THREAD.start()
-            print("heartbeat_handler: started initialization thread")  # for debug
-        # if thread is finished then we will have "ok" in response, and AppAPI will consider that program is ready.
-        r = "init" if MODEL_INIT_THREAD.is_alive() else "ok"
-        print(f"heartbeat_handler: result={r}")  # for debug
-        return r
-
-
-    @APP.on_event("startup")
-    def initialization():
-        # Provide our custom **heartbeat_handler** to set_handlers
-        set_handlers(APP, enabled_handler, heartbeat_handler)
-
-
-.. note:: While this may not be the most ideal way to download models, it remains a viable method.
-    In the future, a more efficient wrapper for model downloading is planned to make the process even more convenient.
-
-Model Storage
-"""""""""""""
-
-By default, models will be downloaded to a directory that's removed when updating the app.
-To persistently store the models even after updates, add the following line to your code:
+This library also provides an additional functionality over this endpoint for easy downloading of models from the `huggingface <https://huggingface.co>`_.
 
 .. code-block::
 
-    from nc_py_api.ex_app import persist_transformers_cache  # noqa # isort:skip
+    @asynccontextmanager
+    async def lifespan(_app: FastAPI):
+        set_handlers(APP, enabled_handler, models_to_fetch=[MODEL_NAME])
+        yield
 
-This will set ``TRANSFORMERS_CACHE`` environment variable to point to the application persistent storage.
-Import of this **must be** on top before importing any code that perform the import of the ``transformers`` library.
+This will automatically download models specified in ``models_to_fetch`` parameter to the application persistent storage.
 
-And that is all, ``transformers`` will automatically download all
-models you use to the **Application Persistent Storage** and AppAPI will keep it between updates.
+If you want write your own logic, you can always pass your own defined ``init_handler`` callback to ``set_handlers``.
 
 Working with Language Models
 """"""""""""""""""""""""""""
@@ -122,13 +79,16 @@ Finally, we arrive at the core aspect of the application, where we interact with
         r = re.search(r"@ai\s(.*)", message.object_content["message"], re.IGNORECASE)
         if r is None:
             return
-        model = pipeline("text2text-generation", model=MODEL_NAME)
+        model = pipeline(
+            "text2text-generation",
+            model=snapshot_download(MODEL_NAME, local_files_only=True, cache_dir=persistent_storage()),
+        )
         # Pass all text after "@ai" we to the Language model.
         response_text = model(r.group(1), max_length=64, do_sample=True)[0]["generated_text"]
         AI_BOT.send_message(response_text, message)
 
 
-Simply put, the AI logic is just two lines of code when using Transformers, which is incredibly efficient and cool.
+Simply put, AI logic is a few lines of code when using Transformers, which is incredibly efficient and cool.
 
 Messages from the AI model are then sent back to Talk Chat as you would expect from a typical chatbot.
 

docs/NextcloudTalkBotTransformers.rst

@@ -24,11 +24,16 @@ build-push:
 	docker login ghcr.io
 	docker buildx build --push --platform linux/arm64/v8,linux/amd64 --tag ghcr.io/cloud-py-api/skeleton:latest .
 
-.PHONY: deploy
-deploy:
+.PHONY: deploy28
+deploy28:
 	docker exec master-nextcloud-1 sudo -u www-data php occ app_api:app:deploy skeleton docker_dev \
 		--info-xml https://raw.githubusercontent.com/cloud-py-api/nc_py_api/main/examples/as_app/skeleton/appinfo/info.xml
 
+.PHONY: deploy27
+deploy27:
+	docker exec master-stable27-1 sudo -u www-data php occ app_api:app:deploy skeleton docker_dev \
+		--info-xml https://raw.githubusercontent.com/cloud-py-api/nc_py_api/main/examples/as_app/skeleton/appinfo/info.xml
+
 .PHONY: run28
 run28:
 	docker exec master-nextcloud-1 sudo -u www-data php occ app_api:app:unregister skeleton --silent || true
@@ -41,9 +46,16 @@ run27:
 	docker exec master-stable27-1 sudo -u www-data php occ app_api:app:register skeleton docker_dev --force-scopes \
 		--info-xml https://raw.githubusercontent.com/cloud-py-api/nc_py_api/main/examples/as_app/skeleton/appinfo/info.xml
 
-.PHONY: manual_register
-manual_register:
+.PHONY: register28
+register28:
 	docker exec master-nextcloud-1 sudo -u www-data php occ app_api:app:unregister skeleton --silent || true
 	docker exec master-nextcloud-1 sudo -u www-data php occ app_api:app:register skeleton manual_install --json-info \
   "{\"appid\":\"skeleton\",\"name\":\"App Skeleton\",\"daemon_config_name\":\"manual_install\",\"version\":\"1.0.0\",\"secret\":\"12345\",\"host\":\"host.docker.internal\",\"port\":9030,\"scopes\":{\"required\":[],\"optional\":[]},\"protocol\":\"http\",\"system_app\":0}" \
   --force-scopes
+
+.PHONY: register27
+register27:
+	docker exec master-stable27-1 sudo -u www-data php occ app_api:app:unregister skeleton --silent || true
+	docker exec master-stable27-1 sudo -u www-data php occ app_api:app:register skeleton manual_install --json-info \
+  "{\"appid\":\"skeleton\",\"name\":\"App Skeleton\",\"daemon_config_name\":\"manual_install\",\"version\":\"1.0.0\",\"secret\":\"12345\",\"host\":\"host.docker.internal\",\"port\":9030,\"scopes\":{\"required\":[],\"optional\":[]},\"protocol\":\"http\",\"system_app\":0}" \
+  --force-scopes

examples/as_app/skeleton/Makefile

@@ -1 +1 @@
-nc_py_api[app]>=0.4.1
+nc_py_api[app]>=0.5.0

examples/as_app/skeleton/requirements.txt

@@ -1,13 +1,19 @@
-"""
-Simplest example.
-"""
+"""Simplest example."""
+from contextlib import asynccontextmanager
 
 from fastapi import FastAPI
 
 from nc_py_api import NextcloudApp
 from nc_py_api.ex_app import LogLvl, run_app, set_handlers
 
-APP = FastAPI()
+
+@asynccontextmanager
+async def lifespan(_app: FastAPI):
+    set_handlers(APP, enabled_handler)
+    yield
+
+
+APP = FastAPI(lifespan=lifespan)
 
 
 def enabled_handler(enabled: bool, nc: NextcloudApp) -> str:
@@ -22,13 +28,6 @@ def enabled_handler(enabled: bool, nc: NextcloudApp) -> str:
     return ""
 
 
-# Of course, you can use `FastAPI lifespan: <https://fastapi.tiangolo.com/advanced/events/#lifespan>` instead of this.
-# The only requirement for the application is to define `/enabled` and `/heartbeat` handlers.
-@APP.on_event("startup")
-def initialization():
-    set_handlers(APP, enabled_handler)
-
-
 if __name__ == "__main__":
     # Wrapper around `uvicorn.run`.
     # You are free to call it directly, with just using the `APP_HOST` and `APP_PORT` variables from the environment.