Add first working version

README.md

@@ -0,0 +1,63 @@
+# shellcheck-gpt
+
+Automatically correct shell script issues by feeding [shellcheck](https://www.shellcheck.net) analysis into an LLM.
+
+## Getting started
+
+### Prerequisites
+
+- [ShellCheck](https://www.shellcheck.net) should be installed and in your $PATH.
+- [Go](https://go.dev/doc/install)
+- A valid OpenAI API key
+
+### Basic installation
+
+shellcheck-gpt can be installed using the `go install` tool. Distribution methods that do not depend on Go are coming soon!
+
+```shell
+$ go install github.com/kkyr/shellcheck-gpt
+```
+
+## Usage
+
+Store your OpenAI API key into the environment:
+
+```shell
+export OPENAI_API_KEY=replace-with-your-key
+```
+
+Run shellcheck-gpt against a script:
+
+```shell
+$ shellcheck-gpt script.sh
+```
+
+This will:
+
+1. Run shellcheck against `script.sh`
+1. Feed the script and the output of shellcheck into an LLM and ask it to make the corrections
+1. Write the LLM's output onto stdout
+
+If you'd like to write the output back into the file, use the `-w` flag:
+
+```shell
+$ shellcheck-gpt -w script.sh
+```
+
+## Configuration
+
+By default, shellcheck-gpt uses the gpt-3.5-turbo model. You can modify this to another model using the `-m` flag:
+
+```shell
+$ shellcheck-gpt -m gpt-4-turbo script.sh
+```
+
+See available options and models using the `--help` flag:
+
+```shell
+$ shellcheck-gpt --help
+```
+
+## Contributing
+
+Contributions are welcome!

example/script.sh

@@ -0,0 +1,7 @@
+#!/bin/sh
+## Example: a typical script with several problems
+for f in *.m3u
+do
+  grep -qi 'hq.*mp3' "$f" \
+    && printf 'Playlist %s contains a HQ file in mp3 format\n' "$f"
+done

go.mod

@@ -0,0 +1,16 @@
+module github.com/kkyr/shellcheck-gpt
+
+go 1.21
+
+require (
+	github.com/briandowns/spinner v1.23.0
+	github.com/fatih/color v1.16.0
+	github.com/sashabaranov/go-openai v1.17.8
+)
+
+require (
+	github.com/mattn/go-colorable v0.1.13 // indirect
+	github.com/mattn/go-isatty v0.0.20 // indirect
+	golang.org/x/sys v0.14.0 // indirect
+	golang.org/x/term v0.1.0 // indirect
+)

go.sum

@@ -0,0 +1,17 @@
+github.com/briandowns/spinner v1.23.0 h1:alDF2guRWqa/FOZZYWjlMIx2L6H0wyewPxo/CH4Pt2A=
+github.com/briandowns/spinner v1.23.0/go.mod h1:rPG4gmXeN3wQV/TsAY4w8lPdIM6RX3yqeBQJSrbXjuE=
+github.com/fatih/color v1.16.0 h1:zmkK9Ngbjj+K0yRhTVONQh1p/HknKYSlNT+vZCzyokM=
+github.com/fatih/color v1.16.0/go.mod h1:fL2Sau1YI5c0pdGEVCbKQbLXB6edEj1ZgiY4NijnWvE=
+github.com/mattn/go-colorable v0.1.13 h1:fFA4WZxdEF4tXPZVKMLwD8oUnCTTo08duU7wxecdEvA=
+github.com/mattn/go-colorable v0.1.13/go.mod h1:7S9/ev0klgBDR4GtXTXX8a3vIGJpMovkB8vQcUbaXHg=
+github.com/mattn/go-isatty v0.0.16/go.mod h1:kYGgaQfpe5nmfYZH+SKPsOc2e4SrIfOl2e/yFXSvRLM=
+github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
+github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
+github.com/sashabaranov/go-openai v1.17.8 h1:snuE7l0XQ1KAmkY/cODAEgxu2fl+g/ybXK6cKQzli/E=
+github.com/sashabaranov/go-openai v1.17.8/go.mod h1:lj5b/K+zjTSFxVLijLSTDZuP7adOgerWeFyZLUhAKRg=
+golang.org/x/sys v0.0.0-20220811171246-fbc7d0a398ab/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.14.0 h1:Vz7Qs629MkJkGyHxUlRHizWJRG2j8fbQKjELVSNhy7Q=
+golang.org/x/sys v0.14.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/term v0.1.0 h1:g6Z6vPFA9dYBAF7DWcH6sCcOntplXsDKcliusYijMlw=
+golang.org/x/term v0.1.0/go.mod h1:jbD1KX2456YbFQfuXm/mYQcufACuNUgVhRMnK/tPxf8=

main.go

@@ -0,0 +1,220 @@
+package main
+
+import (
+	"bytes"
+	"context"
+	"flag"
+	"fmt"
+	"log"
+	"os"
+	"os/exec"
+	"runtime"
+	"text/template"
+	"time"
+
+	"github.com/briandowns/spinner"
+	"github.com/fatih/color"
+	openai "github.com/sashabaranov/go-openai"
+)
+
+const (
+	systemPrompt = `
+	You are a highly skilled systems engineer with expertise in crafting flawless shell scripts. 
+	Your primary responsibility involves two key tasks: Firstly, receiving a shell script that has 
+	been evaluated by a static analysis tool; and secondly, analyzing the provided list of warnings 
+	and issues identified by this tool. Utilizing these inputs, your objective is to develop an 
+	enhanced version of the shell script that effectively addresses and resolves all identified 
+	warnings and issues.
+	
+	Your task is to revise a shell script strictly based on the warnings and errors highlighted by 
+	the static analysis tool. You are required to make changes only to address these specific issues. 
+	Any modifications that do not directly or indirectly relate to correcting these warnings and errors 
+	should be avoided. Additionally, you must preserve the original comments in the script, altering 
+	them only if they become irrelevant due to the changes you implement.
+
+	Your response should exclusively consist of the updated shell script text, presented without using 
+	a code block format.
+	`
+
+	userPrompt = `
+	SHELL_SCRIPT:
+	{{.ScriptContents}}
+	
+	STATIC_ANALYSIS_OUTPUT:
+	{{.StaticAnalysisOutput}}
+	
+	Your task is to revise the provided shell script, focusing solely on rectifying the warnings and errors 
+	identified in the static analysis output. 
+	Ensure that the output of your task is solely the modified shell script text, presented without the use 
+	of a code block format.
+	`
+
+	gpt35turbo = "gpt-3.5-turbo"
+	gpt4turbo = "gpt-4-turbo"
+)
+
+var (
+	userPromptTmpl = template.Must(template.New("prompt").Parse(userPrompt))
+	client = openai.NewClient(os.Getenv("OPENAI_API_KEY")) 
+)
+
+var (
+	// Command line flags
+	writeFile bool
+	showVersion bool
+	model string
+
+	version = "dev"
+)
+
+func init() {
+	flag.BoolVar(&writeFile, "w", false, "write shell script to input file")
+	flag.BoolVar(&showVersion, "v", false, "print version number and exit")
+	flag.StringVar(&model, "m", gpt35turbo, fmt.Sprintf("specify the model to use (%s or %s)", gpt35turbo, gpt4turbo))
+
+	flag.Usage = usage
+}
+
+func usage() {
+	fmt.Fprintf(os.Stderr, "Usage: %s [OPTIONS] FILE\n\n", os.Args[0])
+	fmt.Fprintf(os.Stderr, "Execute shellcheck on the given script and pass the results to a large language model " +
+		"for making appropriate corrections.\n\n")
+	fmt.Fprintf(os.Stderr, "The default behavior displays the modified script on the console. Use the '-w' flag " +
+		"to save the changes directly to the specified file.\n\n")
+		fmt.Fprintf(os.Stderr, "The shellcheck binary must be present in your path.\n\n")
+	fmt.Fprintln(os.Stderr, "OPTIONS:")
+	flag.PrintDefaults()
+	fmt.Fprintln(os.Stderr, "")
+	fmt.Fprintln(os.Stderr, "ENVIRONMENT:")
+	fmt.Fprintln(os.Stderr, "  OPENAI_API_KEY OpenAI API key")
+}
+
+func printf(format string, a ...interface{}) (n int, err error) {
+	return fmt.Fprintf(color.Output, format, a...)
+}
+
+func main() {
+	flag.Parse()
+
+	if showVersion {
+		fmt.Printf("%s %s (runtime: %s)\n", os.Args[0], version, runtime.Version())
+		os.Exit(0)
+	}
+
+	if model != gpt35turbo && model != gpt4turbo {
+		fmt.Fprintf(os.Stderr, "%s: model must be %s or %s\n", os.Args[0], gpt35turbo, gpt4turbo)
+		os.Exit(1)
+	}
+
+	args := flag.Args()
+	if len(args) != 1 {
+		flag.Usage()
+		os.Exit(1)
+	}
+	filePath := args[0]
+
+	run(filePath)
+}
+
+func run(filePath string) {
+	printf("%s\n", "Running shellcheck against script")
+	analysis, err := runShellcheck(filePath)
+	if err != nil {
+		log.Fatal(err)
+	}
+
+	if analysis == "" {
+		printf("%s\n", color.GreenString("shellcheck detected no issues."))
+		return
+	}
+
+	printf("%s\n", color.YellowString("shellcheck detected potential issues"))
+
+	script, err := os.ReadFile(filePath)
+	if err != nil {
+		log.Fatalf("unable to read file: %v", err)
+	}
+
+	result, err := callCompletionAPI(string(script), string(analysis))
+	if err != nil {
+		log.Fatalf("error calling completion API: %v", err)
+	}
+
+	if writeFile {
+		os.WriteFile(filePath, []byte(result), 0644)
+		printf("%s %s\n", color.GreenString("Updated script written to"), color.GreenString(filePath))
+		printf("%s\n", color.YellowString("Double check it before you commit!"))
+	} else {
+		printf("%s\n", result)
+	}
+}
+
+func runShellcheck(filePath string) (string, error) {
+	cmd := exec.Command("shellcheck", filePath)
+
+	output, err := cmd.CombinedOutput()
+	if err != nil {
+		exitCode := cmd.ProcessState.ExitCode()
+		// shellcheck returns exit code 1 when it finds issues
+		if exitCode == 1 {
+			return string(output), nil
+		} else {	
+			return "", fmt.Errorf("shellcheck exited with code %d: %s", exitCode, err)
+		}
+	}
+
+	return "", nil
+}
+
+func callCompletionAPI(script, analysis string) (string, error) {
+	spin := spinner.New(spinner.CharSets[26], 250*time.Millisecond)
+	spin.Prefix = "Waiting for completion response"
+	spin.Start()
+	defer spin.Stop()
+
+	data := map[string]string{
+		"ScriptContents":       string(script),
+		"StaticAnalysisOutput": string(analysis),
+	}
+
+	var buffer bytes.Buffer
+	err := userPromptTmpl.Execute(&buffer, data)
+	if err != nil {
+		log.Fatalf("unable to format prompt: %v", err)
+	}
+
+	resp, err := client.CreateChatCompletion(
+		context.Background(),
+		getCompletionRequest(buffer.String()),
+	)
+	if err != nil {
+		return "", err
+	}
+
+	if len(resp.Choices) > 0 {
+		return resp.Choices[0].Message.Content, nil
+	}
+
+	return "", fmt.Errorf("empty response")
+}
+
+func getCompletionRequest(prompt string) openai.ChatCompletionRequest {
+	m := openai.GPT3Dot5Turbo
+	if model == gpt4turbo {
+		m = openai.GPT4TurboPreview
+	}
+
+	return openai.ChatCompletionRequest{
+		Model: m,
+		Messages: []openai.ChatCompletionMessage{
+			{
+				Role: openai.ChatMessageRoleSystem,
+				Content: systemPrompt,
+			},
+			{
+				Role:    openai.ChatMessageRoleUser,
+				Content: prompt,
+			},
+		},
+	}
+}