From f7b4ffa69b7d4771a3b772170f2439e1f0dfeaae Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 15:37:44 -0800
Subject: [PATCH 01/13] update gitignore

---
 .gitignore | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/.gitignore b/.gitignore
index b6e4761..9b5e917 100644
--- a/.gitignore
+++ b/.gitignore
@@ -127,3 +127,6 @@ dmypy.json
 
 # Pyre type checker
 .pyre/
+
+# Misc / user
+.history

From e134253a83da7e86f005b86ee605f186532db80c Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 15:37:59 -0800
Subject: [PATCH 02/13] update model to mpnet

---
 download.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/download.py b/download.py
index 9f2956d..e848970 100644
--- a/download.py
+++ b/download.py
@@ -7,7 +7,7 @@
 
 def download_model():
     # do a dry run of loading the huggingface model, which will download weights
-    pipeline('fill-mask', model='bert-base-uncased')
+    pipeline('fill-mask', model='sentence-transformers/paraphrase-mpnet-base-v2')
 
 if __name__ == "__main__":
     download_model()
\ No newline at end of file

From 6d9f8e6f024596e8a7ee6e6647b9c4d47a2aa12f Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 16:07:31 -0800
Subject: [PATCH 03/13] replace bert with mpnet

---
 app.py           | 17 ++++++++---------
 requirements.txt |  2 +-
 2 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/app.py b/app.py
index 7f6b061..12387d4 100644
--- a/app.py
+++ b/app.py
@@ -1,24 +1,23 @@
-from transformers import pipeline
-import torch
+from sentence_transformers import SentenceTransformer
+
 
 # Init is ran on server startup
 # Load your model to GPU as a global variable here using the variable name "model"
 def init():
     global model
-    
-    device = 0 if torch.cuda.is_available() else -1
-    model = pipeline('fill-mask', model='bert-base-uncased', device=device)
+    model = SentenceTransformer("sentence-transformers/paraphrase-mpnet-base-v2")
+
 
 # Inference is ran for every server call
 # Reference your preloaded global model variable here.
-def inference(model_inputs:dict) -> dict:
+def inference(model_inputs: dict) -> dict:
     global model
 
     # Parse out your arguments
-    prompt = model_inputs.get('prompt', None)
+    prompt = model_inputs.get("prompt", None)
     if prompt == None:
-        return {'message': "No prompt provided"}
-    
+        return {"message": "No prompt provided"}
+
     # Run the model
     result = model(prompt)
 
diff --git a/requirements.txt b/requirements.txt
index f9cbeac..40b5e65 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,3 @@
 sanic==22.6.2
-transformers
+sentence_transfomers
 accelerate

From 6390b745999c4037a051122e19e94a6d0f7f4285 Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 16:16:53 -0800
Subject: [PATCH 04/13] add specific sentence_transformers version

---
 requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/requirements.txt b/requirements.txt
index 40b5e65..6ee458e 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,3 @@
 sanic==22.6.2
-sentence_transfomers
+sentence_transfomers==2.2.2
 accelerate

From f8edacc4315390546c94aa09295d5b568c8b068a Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 16:22:17 -0800
Subject: [PATCH 05/13] hyphen, not underscore

---
 requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/requirements.txt b/requirements.txt
index 6ee458e..fa57448 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,3 @@
 sanic==22.6.2
-sentence_transfomers==2.2.2
+sentence-transfomers==2.2.2
 accelerate

From 4dd89efe5f7f8f786e7f1c17e5580a2bc6927da6 Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 16:27:03 -0800
Subject: [PATCH 06/13] I am retarded

---
 requirements.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/requirements.txt b/requirements.txt
index fa57448..a80190b 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,3 @@
 sanic==22.6.2
-sentence-transfomers==2.2.2
+sentence-transformers==2.2.2
 accelerate

From f5d3a2ecaa6b2d964cfa0a13c3b5da25c393ac4d Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 16:38:29 -0800
Subject: [PATCH 07/13] add .encode()

---
 app.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/app.py b/app.py
index 12387d4..c333ae3 100644
--- a/app.py
+++ b/app.py
@@ -19,7 +19,7 @@ def inference(model_inputs: dict) -> dict:
         return {"message": "No prompt provided"}
 
     # Run the model
-    result = model(prompt)
+    result = model.encode(prompt)
 
     # Return the results as a dictionary
     return result

From 2efd5aa439bd4dcf41fedd6d50f171f2d4be3ab4 Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 17:31:36 -0800
Subject: [PATCH 08/13] return JSON output

---
 app.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/app.py b/app.py
index c333ae3..2fef3ea 100644
--- a/app.py
+++ b/app.py
@@ -22,4 +22,4 @@ def inference(model_inputs: dict) -> dict:
     result = model.encode(prompt)
 
     # Return the results as a dictionary
-    return result
+    return {"data": result}

From 1fd0f229d025a8843541f1cd3390fb511c6ca989 Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 18:00:21 -0800
Subject: [PATCH 09/13] push a commit for banana re-deploy

---
 app.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/app.py b/app.py
index 2fef3ea..4d4418a 100644
--- a/app.py
+++ b/app.py
@@ -22,4 +22,4 @@ def inference(model_inputs: dict) -> dict:
     result = model.encode(prompt)
 
     # Return the results as a dictionary
-    return {"data": result}
+    return { "data": result }

From 9c5d6b17117f603ca505fbbc9462f7b6d374472f Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Sat, 4 Feb 2023 18:40:35 -0800
Subject: [PATCH 10/13] a python array isn't a list!

---
 app.py | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/app.py b/app.py
index 4d4418a..d23600e 100644
--- a/app.py
+++ b/app.py
@@ -1,4 +1,5 @@
 from sentence_transformers import SentenceTransformer
+from sklearn.preprocessing import normalize
 
 
 # Init is ran on server startup
@@ -19,7 +20,11 @@ def inference(model_inputs: dict) -> dict:
         return {"message": "No prompt provided"}
 
     # Run the model
-    result = model.encode(prompt)
+    sentence_embeddings = model.encode(prompt)
+    normalized_embeddings = normalize(sentence_embeddings)
+
+    # Convert the output array to a list
+    output = normalized_embeddings.tolist()
 
     # Return the results as a dictionary
-    return { "data": result }
+    return { "data": output }

From 31566659bc48f69dc8e66688de8a5fae6dce952c Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Wed, 8 Feb 2023 20:02:54 -0800
Subject: [PATCH 11/13] remove unecessary argument to pipeline in download.py

---
 download.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/download.py b/download.py
index e848970..611be8d 100644
--- a/download.py
+++ b/download.py
@@ -7,7 +7,7 @@
 
 def download_model():
     # do a dry run of loading the huggingface model, which will download weights
-    pipeline('fill-mask', model='sentence-transformers/paraphrase-mpnet-base-v2')
+    pipeline(model='sentence-transformers/paraphrase-mpnet-base-v2')
 
 if __name__ == "__main__":
     download_model()
\ No newline at end of file

From 4a35620b12e9efd69fc557ff1f1b66e71fb6ee2d Mon Sep 17 00:00:00 2001
From: Tyler Termini <69917602+waycroft@users.noreply.github.com>
Date: Wed, 8 Feb 2023 20:12:25 -0800
Subject: [PATCH 12/13] doh

---
 download.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/download.py b/download.py
index 611be8d..d693872 100644
--- a/download.py
+++ b/download.py
@@ -3,11 +3,11 @@
 
 # In this example: A Huggingface BERT model
 
-from transformers import pipeline
+from sentence_transformers import SentenceTransformer
 
 def download_model():
     # do a dry run of loading the huggingface model, which will download weights
-    pipeline(model='sentence-transformers/paraphrase-mpnet-base-v2')
+    SentenceTransformer('sentence-transformers/paraphrase-mpnet-base-v2')
 
 if __name__ == "__main__":
     download_model()
\ No newline at end of file

From a9d1e3c8432e0caff70e798061c506e23b4cd932 Mon Sep 17 00:00:00 2001
From: ramsis <ramsishammadi@gmail.com>
Date: Tue, 4 Jul 2023 20:13:32 +0100
Subject: [PATCH 13/13] migrate to potassium

---
 Dockerfile         | 10 ++++------
 README.md          | 34 ++++++++++++++--------------------
 app.py             | 41 ++++++++++++++++++++++++-----------------
 banana_config.json | 24 ++++++++++++++++++++++++
 requirements.txt   |  2 +-
 server.py          | 42 ------------------------------------------
 test.py            | 10 ----------
 7 files changed, 67 insertions(+), 96 deletions(-)
 create mode 100644 banana_config.json
 delete mode 100644 server.py
 delete mode 100644 test.py

diff --git a/Dockerfile b/Dockerfile
index 888a8da..950601a 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,3 +1,5 @@
+# This is a potassium-standard dockerfile, compatible with Banana
+
 # Must use a Cuda version 11+
 FROM pytorch/pytorch:1.11.0-cuda11.3-cudnn8-runtime
 
@@ -11,18 +13,14 @@ RUN pip3 install --upgrade pip
 ADD requirements.txt requirements.txt
 RUN pip3 install -r requirements.txt
 
-# We add the banana boilerplate here
-ADD server.py .
 
 # Add your model weight files 
 # (in this case we have a python script)
 ADD download.py .
 RUN python3 download.py
 
-
-# Add your custom app code, init() and inference()
-ADD app.py .
+ADD . .
 
 EXPOSE 8000
 
-CMD python3 -u server.py
+CMD python3 -u app.py
\ No newline at end of file
diff --git a/README.md b/README.md
index 89068bd..28815aa 100644
--- a/README.md
+++ b/README.md
@@ -1,20 +1,14 @@
-
-# 🍌 Banana Serverless
-
-This repo gives a framework to serve ML models in production using simple HTTP servers.
-
-# Quickstart
-**[Follow the quickstart guide in Banana's documentation to use this repo](https://docs.banana.dev/banana-docs/quickstart).** 
-
-*(choose "GitHub Repository" deployment method)*
-
-<br>
-
-# Helpful Links
-Understand the 🍌 [Serverless framework](https://docs.banana.dev/banana-docs/core-concepts/inference-server/serverless-framework) and functionality of each file within it.
-
-Generalize this framework to [deploy anything on Banana](https://docs.banana.dev/banana-docs/resources/how-to-serve-anything-on-banana).
-
-<br>
-
-## Use Banana for scale.
+# My Potassium App
+This is a Potassium HTTP server, created with `banana init` CLI
+
+### Testing
+Start a local dev server with `banana dev`
+
+### Deployment
+1. Create empty repo on [Github](https://github.com)
+2. Push this repo to github
+```
+git remote add origin https://github.com/{username}/{repo-name}.git
+```
+3. [Log into Banana](https://app.banana.dev/onboard)
+4. Select this repo to build and deploy!
\ No newline at end of file
diff --git a/app.py b/app.py
index d23600e..e21af78 100644
--- a/app.py
+++ b/app.py
@@ -1,24 +1,26 @@
+from potassium import Potassium, Request, Response
+
 from sentence_transformers import SentenceTransformer
 from sklearn.preprocessing import normalize
 
+app = Potassium("my_app")
 
-# Init is ran on server startup
-# Load your model to GPU as a global variable here using the variable name "model"
+# @app.init runs at startup, and loads models into the app's context
+@app.init
 def init():
-    global model
     model = SentenceTransformer("sentence-transformers/paraphrase-mpnet-base-v2")
-
-
-# Inference is ran for every server call
-# Reference your preloaded global model variable here.
-def inference(model_inputs: dict) -> dict:
-    global model
-
-    # Parse out your arguments
-    prompt = model_inputs.get("prompt", None)
-    if prompt == None:
-        return {"message": "No prompt provided"}
-
+   
+    context = {
+        "model": model
+    }
+
+    return context
+
+# @app.handler runs for every call
+@app.handler()
+def handler(context: dict, request: Request) -> Response:
+    prompt = request.json.get("prompt")
+    model = context.get("model")
     # Run the model
     sentence_embeddings = model.encode(prompt)
     normalized_embeddings = normalize(sentence_embeddings)
@@ -26,5 +28,10 @@ def inference(model_inputs: dict) -> dict:
     # Convert the output array to a list
     output = normalized_embeddings.tolist()
 
-    # Return the results as a dictionary
-    return { "data": output }
+    return Response(
+        json = {"data": output}, 
+        status=200
+    )
+
+if __name__ == "__main__":
+    app.serve()
\ No newline at end of file
diff --git a/banana_config.json b/banana_config.json
new file mode 100644
index 0000000..cd9edbe
--- /dev/null
+++ b/banana_config.json
@@ -0,0 +1,24 @@
+{
+    "name": "",
+    "category": "",
+    "example_input": {
+        "prompt": "Hello I am a [MASK] model."
+    },
+    "example_output": {
+        "outputs":[
+            {
+                "score":0.13177461922168732,
+                "token":4827,
+                "token_str":"fashion",
+                "sequence":"hello i am a fashion model."
+            },
+            {
+                "score":0.1120428815484047,
+                "token":2535,
+                "token_str":"role",
+                "sequence":"hello i am a role model."
+            }
+        ]
+    },
+    "version": "1"
+}
\ No newline at end of file
diff --git a/requirements.txt b/requirements.txt
index a80190b..10f94c7 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,3 @@
-sanic==22.6.2
+potassium
 sentence-transformers==2.2.2
 accelerate
diff --git a/server.py b/server.py
deleted file mode 100644
index d66cf1a..0000000
--- a/server.py
+++ /dev/null
@@ -1,42 +0,0 @@
-# Do not edit if deploying to Banana Serverless
-# This file is boilerplate for the http server, and follows a strict interface.
-
-# Instead, edit the init() and inference() functions in app.py
-
-from sanic import Sanic, response
-import subprocess
-import app as user_src
-
-# We do the model load-to-GPU step on server startup
-# so the model object is available globally for reuse
-user_src.init()
-
-# Create the http server app
-server = Sanic("my_app")
-
-# Healthchecks verify that the environment is correct on Banana Serverless
-@server.route('/healthcheck', methods=["GET"])
-def healthcheck(request):
-    # dependency free way to check if GPU is visible
-    gpu = False
-    out = subprocess.run("nvidia-smi", shell=True)
-    if out.returncode == 0: # success state on shell command
-        gpu = True
-
-    return response.json({"state": "healthy", "gpu": gpu})
-
-# Inference POST handler at '/' is called for every http call from Banana
-@server.route('/', methods=["POST"]) 
-def inference(request):
-    try:
-        model_inputs = response.json.loads(request.json)
-    except:
-        model_inputs = request.json
-
-    output = user_src.inference(model_inputs)
-
-    return response.json(output)
-
-
-if __name__ == '__main__':
-    server.run(host='0.0.0.0', port=8000, workers=1)
diff --git a/test.py b/test.py
deleted file mode 100644
index 3c88413..0000000
--- a/test.py
+++ /dev/null
@@ -1,10 +0,0 @@
-# This file is used to verify your http server acts as expected
-# Run it with `python3 test.py``
-
-import requests
-
-model_inputs = {'prompt': 'Hello I am a [MASK] model.'}
-
-res = requests.post('http://localhost:8000/', json = model_inputs)
-
-print(res.json())
\ No newline at end of file