bin: add a script for generating the api client

Stonelinks · Stonelinks · commit 698d043e90ed · 2023-05-02T16:56:59.000-07:00
I tried a bunch of generators - they all had issues. Some generated thousands of lines of unintelligible code. Others generated code that had errors, failed because of our spec or made weird assumptions about library structure. Others I didn't even bother with since they seemed abandoned. I eventually settled on https://github.com/Azure/autorest by Microsoft. Its got some things going for it: - It works! - Its tested and documented - It gives decent configuration options for the generator - It allows for decent customization by users and consumers of the client - It can (eventually) support streaming responses from the server - The only dependency is docker - Its actually supported by a large organization / number of developers - It could be trivially extended to generate clients for multiple languages besides python The downsides: - Its still generating a lot of code that isn't very idiomatic python (but its still reasonable to follow) - It depends on some Microsoft / Azure libraries for basic things instead... they're probably fine but its still a bit weird to be using msrest and azure-core instead of like... pydantic and requests. - The docker container is slow to start since it always downloads the latest version of autorest Overall the pros outweigh the cons IMO... Going to go forward with this.
diff --git a/bin/generate-api-client b/bin/generate-api-client
@@ -0,0 +1,66 @@
+#!/bin/bash
+set -e
+
+cd $(dirname "$0")/../
+
+CLIENT_PACKAGE_NAME=llama_cpp.client
+CLIENT_TMP_DIR=.tmp/llama-cpp-openapi-client
+CLIENT_FINAL_DIR=llama_cpp/client
+
+# Ensure we're starting from a clean slate
+rm -rf .tmp $CLIENT_TMP_DIR $CLIENT_FINAL_DIR
+
+# Pull the latest version of the API's OpenAPI spec
+./bin/save-openapi-json
+OPENAPI_JSON_PATH=.tmp/llama_server_openapi.json
+
+# Generate the client
+
+# Version tolerant: this is apparently their newer, more pythonic generator that is more tolerant to changes in the API
+VERSION_TOLERANT=true
+
+# Python3 only: we don't need to support python2
+PYTHON3_ONLY=true
+
+# Low level client: I think this is yet another client generator (???) - turn it off for now
+LOW_LEVEL_CLIENT=false
+
+# Models: IMO the client it totally pointless without generating models (which is disabled by default) - so enable it
+# The only models that are supported are for msrest
+MODELS_MODE=msrest
+
+# Client-side validation: also something disabled by default - i'm going to turn this off for now because i think there might be some problems with return types in the API
+CLIENT_SIDE_VALIDATION=false
+
+# Setup.py: the client lives in llama_cpp/client, don't generate a setup.py or anything else to support a standalone package
+BASIC_SETUP_PY=false
+
+# RXun autorest
+docker run --rm \
+    -v $PWD:/local \
+    -w /local \
+    azsdkengsys.azurecr.io/azuresdk/autorest-python \
+    autorest \
+    --python \
+    --namespace=$CLIENT_PACKAGE_NAME \
+    --package-name=llama-cpp-api-client \
+    --package-version=0.0.1 \
+    --basic-setup-py=$BASIC_SETUP_PY \
+    --version-tolerant=$VERSION_TOLERANT \
+    --python3-only=$PYTHON3_ONLY \
+    --low-level-client=$LOW_LEVEL_CLIENT \
+    --models-mode=$MODELS_MODE \
+    --client-side-validation=$CLIENT_SIDE_VALIDATION \
+    --combine-operation-files=false \
+    --clear-output-folder=true \
+    --no-namespace-folders=true \
+    --input-file=/local/$OPENAPI_JSON_PATH \
+    --output-folder=/local/$CLIENT_TMP_DIR
+
+# Move the client to the final location
+cp -r $CLIENT_TMP_DIR $CLIENT_FINAL_DIR
+
+# Do some cleanup with black, isort and autoflake
+black $CLIENT_FINAL_DIR
+isort $CLIENT_FINAL_DIR
+autoflake -r --in-place --remove-all-unused-imports --remove-unused-variables $CLIENT_FINAL_DIR
