Skip to content
This repository has been archived by the owner on Nov 16, 2024. It is now read-only.
/ dovelet Public archive
forked from kaneshin/pigeon

A Google Cloud Vision tool and library wrapper.

License

Notifications You must be signed in to change notification settings

EdoardoLaGreca/dovelet

 
 

Repository files navigation

Dovelet

Dovelet is both a Go wrapper library for the Google Cloud Vision API and a tool to make queries using such wrapper.

Go workflow status Go Report Card

The original Pigeon tool and library used a deprecated Google Cloud Vision API. One reason could be that, at the time of writing, its last commit was on Oct 14 2020, which means that the API may have become deprecated somewhere after the last commit. This fork attempts to stay up to date and to pursue a similar purpose. Apart from the usage of deprecated APIs, I think that the original Pigeon codebase was unnecessarily complex. This fork has been renamed to distinguish it from the original.

Requirements

You need to create a shell variable named GOOGLE_APPLICATION_CREDENTIALS which holds the path of your Service Account private key (replace /path/to/service_account.json). The key type must be JSON.

GOOGLE_APPLICATION_CREDENTIALS=/path/to/service_account.json
export GOOGLE_APPLICATION_CREDENTIALS

To generate the credentials file refer to "Authentication with service accounts" and "Create a service account key".

The dovelet tool

The dovelet executable binary is a command-line tool to interface with the Google Cloud Vision API.

Installation

Using go install

go install github.com/EdoardoLaGreca/dovelet/cmd/dovelet@latest

Building from source

git clone https://github.com/EdoardoLaGreca/dovelet
cd dovelet
go install ./cmd/dovelet # or: go build ./cmd/dovelet

Usage

Use the dovelet command to make requests. The syntax is as follows.

dovelet -text [ -lang l1,l2,... ] files ...
dovelet -doc [ -lang l1,l2,... ] files ...
dovelet -face files ...
dovelet -image-properties files ...
dovelet -label files ...
dovelet -landmark files ...
dovelet -logo files ...
dovelet -safe-search files ...

Where:

  • l1,l2,... are language hints for better accuracy (see "languageHints code" in Supported languages)
  • files ... is one or more files to use in the request

You can find more details using dovelet -h.

The dovelet package

The dovelet package contains types and functions for simple Cloud Vision queries.

Installation

go get https://github.com/EdoardoLaGreca/dovelet

Usage

Refer to cmd/dovelet/main.go for an example.

Examples

Pigeon

input:

pigeon

dovelet -label assets/pigeon.png

output:

[
  {
    "label_annotations": [
      {
        "mid": "/m/015p6",
        "description": "Bird",
        "score": 0.96299416,
        "topicality": 0.96299416
      },
      {
        "mid": "/m/01c09k",
        "description": "Stock dove",
        "score": 0.8821084,
        "topicality": 0.8821084
      },
      {
        "mid": "/m/01c4rd",
        "description": "Beak",
        "score": 0.87776655,
        "topicality": 0.87776655
      },
      {
        "mid": "/m/036k5h",
        "description": "Grey",
        "score": 0.8444576,
        "topicality": 0.8444576
      },
      {
        "mid": "/m/01c08f",
        "description": "Rock dove",
        "score": 0.83560807,
        "topicality": 0.83560807
      },
      {
        "mid": "/m/01k0mv",
        "description": "Road surface",
        "score": 0.8078585,
        "topicality": 0.8078585
      },
      {
        "mid": "/m/0dfgf",
        "description": "Feather",
        "score": 0.8044634,
        "topicality": 0.8044634
      },
      {
        "mid": "/m/083s2",
        "description": "Wing",
        "score": 0.74377084,
        "topicality": 0.74377084
      },
      {
        "mid": "/m/07sx2n",
        "description": "Natural material",
        "score": 0.6731261,
        "topicality": 0.6731261
      },
      {
        "mid": "/m/0hr8",
        "description": "Asphalt",
        "score": 0.6372245,
        "topicality": 0.6372245
      }
    ]
  }
]

Lenna

input:

lenna

dovelet -safe-search assets/lenna.jpg

output:

[
  {
    "safe_search_annotation": {
      "adult": 2,
      "spoof": 1,
      "medical": 2,
      "violence": 1,
      "racy": 4
    }
  }
]

License

MIT

Credits