horus

horus

Overview

horus is a Go error‐handling toolkit that does more than “return an error”

horus captures the what (operation), why (user‐friendly message), where (stack trace), and how (underlying cause and arbitrary key/value details), then lets you:

• Wrap and re‐wrap errors with layered context
• Seamlessly propagate or bail out with CheckErr (colored/JSON + configurable exit codes)
• Plug into any CLI, HTTP handler, or logger with zero ceremony
• Whether you’re building a command‐line tool, a microservice, or a big data pipeline, horus ensures nothing gets lost in translation when things go sideways

Features

Context-Rich Errors

Create Herror instances via NewHerror, NewCategorizedHerror, Wrap or WithDetail that carry:

• Operation name (Op)
• Human-readable message (Message)
• Category tag (Category)
• Arbitrary details map (Details)
• Full stack trace

Error Propagation & Inspection

• PropagateErr for idiomatic upstream wrapping
• RootCause(err) to peel back nested failures
• Helpers like AsHerror, IsHerror, Operation, UserMessage, GetDetail, Category, StackTrace

err := doSomething()
// only if err != nil it gets wrapped
return horus.PropagateErr("DoSomething", "SERVICE", "failed", err, nil)

Flexible Formatting

• JSONFormatter for structured logs
• PseudoJSONFormatter for aligned, colorized tables in your terminal
• PlainFormatter or SimpleColoredFormatter for minimal output

Check & Exit

• CheckErr(err, opts...) writes formatted error to your choice of io.Writer and exits with a customizable code
• Built-in overrides: writer, formatter, exit code, operation, category, message, details

horus.CheckErr(err)               // default: colored table + os.Exit(1)
horus.CheckErr(err, horus.WithWriter(os.Stdout), horus.WithExitCode(42))

Not-Found Hooks

• LogNotFound / NullAction implement NotFoundAction for pluggable “resource missing” behaviors
• Fully testable via WithLogWriter

act := horus.LogNotFound("cache miss")
resolved, err := act("user:123")
// resolved==false, err==nil

Test Utilities

• CollectingError (implements io.Writer + error) to capture and inspect output in tests
• Easy use of WithWriter(buf) to drive deterministic output

Panic Integration

• Panic(op, msg) logs a colored panic banner, captures a stack, then panics with a full Herror payload

Quickstart

package main

import (
  "errors"
  "fmt"

  "github.com/DanielRivasMD/horus"
)

func loadConfig(path string) error {
  // pretend this fails
  return errors.New("file not found")
}

func main() {
  err := loadConfig("/etc/app.cfg")
  if err != nil {
    // wrap with context, category, and detail
    wrapped := horus.NewCategorizedHerror(
      "LoadConfig",
      "IO_ERROR",
      "unable to load configuration",
      err,
      map[string]any{"path": "/etc/app.cfg"},
    )
    // print a pretty, colored table to stderr and exit
    horus.CheckErr(wrapped)
  }

  fmt.Println("config loaded")
}

Installation

Language-Specific

Usage

import "github.com/DanielRivasMD/horus"

Error Handling Integration

The horus error-handling library provides a set of powerful functions to wrap, propagate, log, and format errors across your application. Here’s how you can leverage these functions at various layers:

1. Lower-Level Functions

Wrap errors as soon as they occur

For example, when reading a configuration file:

package fileutils

import (
  "fmt"
  "os"

  "github.com/DanielRivasMD/horus"
)

// ReadConfig tries to read a JSON config from disk.
// On failure it wraps the underlying error with full context, category and details.
func ReadConfig(path string) ([]byte, error) {
  data, err := os.ReadFile(path)
  if err != nil {
    return nil, horus.PropagateErr(
      "ReadConfig",                              // Op
      "IO_ERROR",                                // Category
      fmt.Sprintf("unable to load config"),      // Message
      err,                                       // underlying error
      map[string]any{                            // Details
        "path": path,
      },
    )
  }
  return data, nil
}

What this does:

• Uses Go 1.16+’s os.ReadFile instead of the deprecated ioutil
• Always returns nil or a rich *Herror, never a raw error
• Stamps on:
  • Op = "ReadConfig"
  • Category = "IO_ERROR"
  • Message = a user-friendly "unable to load config"
  • Details = {"path": path}
  • Stack trace (captured at the call site)

You’ll now get output like:

Op       ReadConfig,
Message  unable to load config,
Err      open /etc/app.cfg: no such file or directory,
path     /etc/app.cfg,
Category IO_ERROR,

Stack
  fileutils.ReadConfig()
    /Users/.../fileutils/config.go:12
  main.main()
    /Users/.../cmd/app/main.go:23
  runtime.main()
    /usr/local/go/src/runtime/proc.go:250
  ...

and if you prefer JSON:

horus.CheckErr(err, horus.WithFormatter(horus.JSONFormatter))

will emit something like:

{
  "Op": "ReadConfig",
  "Message": "unable to load config",
  "Err": "open /etc/app.cfg: no such file or directory",
  "Details": { "path": "/etc/app.cfg" },
  "Category": "IO_ERROR",
  "Stack": [ ... ]
}

2. Business Logic Functions

When higher-level functions catch errors from lower‐level routines, add domain‐specific context with PropagateErr (or WithDetail) so every layer contributes its own clues:

package business

import (
  "github.com/your_module/fileutils"
  "github.com/DanielRivasMD/horus"
)

// LoadAndProcessConfig orchestrates reading + validating your config.
// Any I/O or parse failures get wrapped with step-specific context.
func LoadAndProcessConfig(configPath string) error {
  // 1. Call the fileutils helper
  data, err := fileutils.ReadConfig(configPath)
  if err != nil {
    // PropagateErr merges the underlying category/details and stamps on new ones.
    return horus.PropagateErr(
      "LoadAndProcessConfig",           // operation name
      "CONFIG_ERROR",                   // business category
      "unable to load application config", // user-friendly message
      err,                              // the error from ReadConfig
      map[string]any{                   // extra details
        "path":   configPath,
        "service": "business",
      },
    )
  }

  // 2. (Optional) add validation context
  if len(data) == 0 {
    // NewHerror creates a fresh Herror; WithDetail would wrap an existing one.
    return horus.NewHerror(
      "LoadAndProcessConfig",
      "config data is empty",
      nil,
      map[string]any{"path": configPath},
    )
  }

  // 3. process the config...
  return nil
}

• PropagateErr automatically carries forward any category or details from the lower layer (and merges the new payload)
• We choose a “business” category ("CONFIG_ERROR") that’s orthogonal to the lower-level "IO_ERROR"
• We include both path and our own service:"business" detail.
• For purely business‐rule failures (like empty data), we use NewHerror to start a fresh error.

3. Centralized Error Reporting and Logging

At the top‐level ofthe app - usually in main() - CheckErr can be use as one‐stop fatal error handler:

• Format the error (colored table by default) or JSON if you prefer
• Register the error’s category in a global registry (for metrics/observability)
• Exit with a configurable code (default: 1)

package main

import (
  "os"

  "github.com/DanielRivasMD/horus"
  "github.com/your_module/business"
)

func main() {
  // Run your business logic
  err := business.LoadAndProcessConfig("config.json")
  if err != nil {
    // Default: colored table → stderr, exit code 1
    horus.CheckErr(err)

    // Or JSON + code 2 + log to stdout:
    // horus.CheckErr(
    //   err,
    //   horus.WithFormatter(horus.JSONFormatter),
    //   horus.WithExitCode(2),
    //   horus.WithWriter(os.Stdout),
    // )
  }

  // Continue with normal execution...
}

Under the hood, CheckErr does:

• RegisterError(err) – increments a counter for your error’s category
• fmt.Fprintln(writer, formatter(err)) – prints your chosen format
• exitFunc(code) – calls os.Exit(code) by default

This ensures that all unhandled, fatal errors flow through a consistent, observable pipeline

4. Integration with External Processes

For commands executed via external processes (e.g., running system commands), use functions like ExecCmd or CaptureExecCmd, horus:

• Shows both ExecCmd (streams output) and CaptureExecCmd (buffers it)
• Wraps errors with PropagateErr for context before calling CheckErr
• Logs stdout/stderr details in your error’s Details map

package domovoi

import (
  "fmt"

  "github.com/DanielRivasMD/horus"
)

// ListDirectory runs `ls -la <path>` twice: once streamed, once captured.
// It returns an error if anything fails; caller can then CheckErr(err).
func ListDirectory(path string) error {
  // 1) Stream mode
  if err := ExecCmd("ls", "-la", path); err != nil {
    // ExecCmd already wraps in *Herror, but we can add our own op/category
    return horus.NewCategorizedHerror(
      "ListDirectory",                // Op
      "SYS_CMD",                      // Category
      fmt.Sprintf("ls -la %s", path), // Message
      err,
      nil, // no extra details here
    )
  }

  // 2) Capture mode
  stdout, stderr, err := CaptureExecCmd("ls", "-la", path)
  if err != nil {
    // We got a *Herror from CaptureExecCmd—merge in the captured output
    return horus.PropagateErr(
      "ListDirectory",
      "SYS_CMD",
      "failed to capture ls output",
      err,
      map[string]any{"stdout": stdout, "stderr": stderr},
    )
  }

  fmt.Println("=== STDOUT ===")
  fmt.Print(stdout)
  fmt.Println("=== STDERR ===")
  fmt.Print(stderr)
  return nil
}

• ExecCmd(op, category, message, *exec.Cmd) runs and logs failures immediately
• CaptureExecCmd returns (stdout, stderr string, err error)
• We propagate errors with PropagateErr so our top‐level CheckErr shows the full story: operation, category, message, underlying cause, stdout, stderr, and stack trace

5. OS Level Operations

When wrapping system calls like os.Chdir, use Horus to enrich and propagate errors with full context:

package domovoi

import (
  "fmt"
  "os"

  "github.com/DanielRivasMD/horus"
)

// ChangeDirectory attempts to chdir into the given path.
// On failure it wraps the underlying os.Chdir error with operation,
// category, user-friendly message, and the path in Details.
func ChangeDirectory(path string) error {
  if err := os.Chdir(path); err != nil {
    return horus.PropagateErr(
      "ChangeDirectory",                               // Op
      "FS_ERROR",                                      // Category
      fmt.Sprintf("unable to change working directory"),// Message
      err,                                             // underlying error
      map[string]any{"path": path},                    // Details
    )
  }
  return nil
}

• Op = "ChangeDirectory"
• Category = "FS_ERROR"
• Message = "unable to change working directory"
• Details = {"path": "/some/dir"}
• Stack trace captured at the call site

That way, any failure bubbles up as a full *Herror - complete with stack, category, and details—making your logs and CLI output immediately actionable

Development

Build from source:

git clone https://github.com/DanielRivasMD/horus
cd horus

Language-Specific Setup

License

Copyright (c) 2025

See the LICENSE file for license details.