alt_text 0.1.1 → 0.3.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (7) hide show
  1. checksums.yaml +4 -4
  2. data/README.md +17 -5
  3. data/bin/check_dependencies.rb +7 -0
  4. data/lib/alt_text/client.rb +48 -24
  5. data/lib/alt_text/llm_registry.rb +6 -4
  6. data/lib/alt_text/version.rb +1 -1
  7. metadata +24 -9
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: a2ce46a5e7a525ed907991d2d50e35d8a2c002f73118e5502e7fe9305775d310
4
- data.tar.gz: 564f906bdcebd926b85429d174e2c4cae9c66b1349973d7ec905cf6c74354bed
3
+ metadata.gz: 3756a52c52a6e8ed15b67514e76dce4424ca4f5a58a948fbec4f34635f800fe9
4
+ data.tar.gz: daee67097afb3acae861cd4b302d4a0c6f85f53fa77e21774eb109a9b9bb0555
5
5
  SHA512:
6
- metadata.gz: 70405efc438ae7ebb3b4ecf7cd4a9225256c44008c9dbed79958381ff48a13fc703cb73ba06c64387f053aa823139de47b18d12500b47c6c18211160d1691ac0
7
- data.tar.gz: 56c6673daa0095dacc08edec163ac82ec9bae3de884e77f263b6b68ecf60501d439a9f7e1a0f25d3156fc91d6d0e61911ea316affc80d580db7b1652f87501b9
6
+ metadata.gz: 65f543a46249258775f43f90d72e2ae1262d288412729f5ca3dc454eea3dcec7c14f1e1f51a23f02010e1a4fb8ce1273f88a5f34ad5a8b27f422b4712a51f6e7
7
+ data.tar.gz: 5d32ebae055a9c93d73731dc435e94e0102219b2c466b3d78486b34c9e126409a50174cca776e9c2cedacd2cd537dc19cd9f34f38f3bbfb1ef5d80584c4c22ac
data/README.md CHANGED
@@ -2,6 +2,8 @@
2
2
 
3
3
  This uses Ruby's AWS SDK to send images and a prompt to an LLM in Amazon's Bedrock to generate Alt Text for the images.
4
4
 
5
+ The client uses Bedrock's `converse` API and currently supports JPEG and PNG inputs.
6
+
5
7
  ### Ruby Client Usage
6
8
 
7
9
  This gem uses imagemagick to resize large images, so you will need to install imagemagick:
@@ -39,19 +41,29 @@ gem install alt_text
39
41
  Instantiate the client with injected AWS credentials:
40
42
 
41
43
  ```
42
- client = AltText::Client.new {
43
- access_key_id: ENV['YOUR_ACCESS_KEY_ID'],
44
- secret_access_key: ENV['YOUR_SECRET_ACCESS_KEY'],
44
+ client = AltText::Client.new(
45
+ access_key: ENV['AWS_ACCESS_KEY_ID'],
46
+ secret_key: ENV['AWS_SECRET_ACCESS_KEY'],
45
47
  region: 'us-east-1'
46
- }
48
+ )
47
49
  ```
48
50
 
49
51
  Call the `#process_image` method with the image path, prompt, and LLM ID as arguments:
50
52
 
51
53
  ```
52
- client.process_image('folder/image.png', 'Please generate alt text', 'sonnet3.51`)
54
+ client.process_image(
55
+ 'folder/image.png',
56
+ prompt: 'Please generate alt text',
57
+ model_id: 'default'
58
+ )
53
59
  ```
54
60
 
61
+ Supported image types:
62
+
63
+ - `.jpg`
64
+ - `.jpeg`
65
+ - `.png`
66
+
55
67
  *Note: A sample prompt can be found in `prompt.txt`.*
56
68
 
57
69
  ### CLI Usage
data/bin/check_dependencies.rb ADDED
@@ -0,0 +1,7 @@
1
+ # frozen_string_literal: true
2
+
3
+ spec = Gem::Specification.load("alt_text.gemspec")
4
+
5
+ spec.runtime_dependencies.each do |dependency|
6
+ system("gem", "list", "--remote", "--exact", dependency.name)
7
+ end
data/lib/alt_text/client.rb CHANGED
@@ -2,7 +2,7 @@
2
2
 
3
3
  require 'aws-sdk-bedrockruntime'
4
4
  require 'mini_magick'
5
- require 'base64'
5
+ require 'marcel'
6
6
 
7
7
  module AltText
8
8
  class Client
@@ -14,38 +14,62 @@ module AltText
14
14
  )
15
15
  end
16
16
 
17
- def process_image(image_path, prompt:, model_id:)
17
+ def process_image(image_path, prompt:, model_id:, temperature: 0.0)
18
18
  model_id = AltText::LLMRegistry.resolve(model_id)
19
+ image_format = image_format_for(image_path)
19
20
  tmp_image = resize_if_needed(image_path)
20
21
 
21
- encoded_image = Base64.strict_encode64(File.binread(tmp_image))
22
+ image_bytes = File.binread(tmp_image)
22
23
  tmp_image.close! if tmp_image.is_a?(Tempfile)
23
24
 
24
- payload = {
25
- messages: [
26
- { role: 'user',
27
- content: [
28
- { type: 'image',
29
- source:
30
- { type: 'base64',
31
- media_type: 'image/jpeg',
32
- data: encoded_image } },
33
- { type: 'text',
34
- text: prompt }
35
- ] }
36
- ],
37
- max_tokens: 10_000,
38
- anthropic_version: 'bedrock-2023-05-31'
39
- }
40
-
41
- response = @client.invoke_model(model_id: model_id,
42
- content_type: 'application/json',
43
- body: payload.to_json)
44
- JSON.parse(response.body.read)['content'][0]['text']
25
+ messages = [
26
+ {
27
+ role: 'user',
28
+ content: [
29
+ {
30
+ image: {
31
+ format: image_format,
32
+ source: {
33
+ bytes: image_bytes
34
+ }
35
+ }
36
+ },
37
+ {
38
+ text: prompt
39
+ }
40
+ ]
41
+ }
42
+ ]
43
+
44
+ # The `converse` method of the Bedrock Ruby SDK is used to interact with
45
+ # LLM models in a standardized way, using a "messages" schema that supports
46
+ # text, images, and tool calls. Unlike `invoke_model`, which requires
47
+ # model-specific payloads. Note that this prevents fine-grained control
48
+ # of image processing parameters that some models may support.
49
+ #
50
+ # Examples of supported models:
51
+ # - Amazon Nova Pro (supports text and images)
52
+ # - Amazon Nova Lite (supports text and images)
53
+ # - Anthropic Claude / Opus (supports text and images)
54
+ response = @client.converse(model_id: model_id,
55
+ messages: messages,
56
+ inference_config: { temperature: temperature })
57
+
58
+ response.output.message.content.first.text
45
59
  end
46
60
 
47
61
  private
48
62
 
63
+ def image_format_for(path)
64
+ content_type = Marcel::MimeType.for(Pathname.new(path))
65
+ case content_type
66
+ when 'image/jpeg' then 'jpeg'
67
+ when 'image/png' then 'png'
68
+ else
69
+ raise ArgumentError, "Unsupported image type: #{content_type || 'unknown'}"
70
+ end
71
+ end
72
+
49
73
  def resize_if_needed(file)
50
74
  if File.size(file) < 4_000_000
51
75
  file
data/lib/alt_text/llm_registry.rb CHANGED
@@ -1,12 +1,14 @@
1
1
  # frozen_string_literal: true
2
2
 
3
3
  module AltText
4
+ # Bedrock model IDs evolve over time, so this mapping is intentionally
5
+ # centralized and easy to update as models are added, renamed, or retired.
6
+ # Entries here are expected to work with the Bedrock Ruby SDK `converse` API.
4
7
  class LLMRegistry
5
8
  LLM_MAP = {
6
- 'default' => 'anthropic.claude-3-5-sonnet-20240620-v1:0',
7
- 'sonnet3.51' => 'anthropic.claude-3-5-sonnet-20240620-v1:0',
8
- 'sonnet3.52' => 'anthropic.claude-3-5-sonnet-20241022-v2:0',
9
- 'sonnet3.571' => 'us.anthropic.claude-3-7-sonnet-20250219-v1:0',
9
+ 'default' => 'us.amazon.nova-pro-v1:0',
10
+ 'novalite' => 'amazon.nova-lite-v1:0',
11
+ 'sonnet4.5' => 'anthropic.claude-sonnet-4-5',
10
12
  'novapro' => 'us.amazon.nova-pro-v1:0'
11
13
  }.freeze
12
14
 
data/lib/alt_text/version.rb CHANGED
@@ -1,5 +1,5 @@
1
1
  # frozen_string_literal: true
2
2
 
3
3
  module AltText
4
- VERSION = '0.1.1'
4
+ VERSION = '0.3.0'
5
5
  end
metadata CHANGED
@@ -1,13 +1,13 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: alt_text
3
3
  version: !ruby/object:Gem::Version
4
- version: 0.1.1
4
+ version: 0.3.0
5
5
  platform: ruby
6
6
  authors:
7
7
  - Alex Kiessling
8
8
  bindir: bin
9
9
  cert_chain: []
10
- date: 2025-11-18 00:00:00.000000000 Z
10
+ date: 1980-01-02 00:00:00.000000000 Z
11
11
  dependencies:
12
12
  - !ruby/object:Gem::Dependency
13
13
  name: aws-sdk-bedrockruntime
@@ -15,42 +15,56 @@ dependencies:
15
15
  requirements:
16
16
  - - "~>"
17
17
  - !ruby/object:Gem::Version
18
- version: 1.55.0
18
+ version: 1.85.0
19
19
  type: :runtime
20
20
  prerelease: false
21
21
  version_requirements: !ruby/object:Gem::Requirement
22
22
  requirements:
23
23
  - - "~>"
24
24
  - !ruby/object:Gem::Version
25
- version: 1.55.0
25
+ version: 1.85.0
26
26
  - !ruby/object:Gem::Dependency
27
27
  name: dotenv
28
28
  requirement: !ruby/object:Gem::Requirement
29
29
  requirements:
30
30
  - - "~>"
31
31
  - !ruby/object:Gem::Version
32
- version: 3.1.8
32
+ version: 3.2.0
33
33
  type: :runtime
34
34
  prerelease: false
35
35
  version_requirements: !ruby/object:Gem::Requirement
36
36
  requirements:
37
37
  - - "~>"
38
38
  - !ruby/object:Gem::Version
39
- version: 3.1.8
39
+ version: 3.2.0
40
+ - !ruby/object:Gem::Dependency
41
+ name: marcel
42
+ requirement: !ruby/object:Gem::Requirement
43
+ requirements:
44
+ - - "~>"
45
+ - !ruby/object:Gem::Version
46
+ version: 2.1.0
47
+ type: :runtime
48
+ prerelease: false
49
+ version_requirements: !ruby/object:Gem::Requirement
50
+ requirements:
51
+ - - "~>"
52
+ - !ruby/object:Gem::Version
53
+ version: 2.1.0
40
54
  - !ruby/object:Gem::Dependency
41
55
  name: mini_magick
42
56
  requirement: !ruby/object:Gem::Requirement
43
57
  requirements:
44
58
  - - "~>"
45
59
  - !ruby/object:Gem::Version
46
- version: 5.3.0
60
+ version: 5.4.0
47
61
  type: :runtime
48
62
  prerelease: false
49
63
  version_requirements: !ruby/object:Gem::Requirement
50
64
  requirements:
51
65
  - - "~>"
52
66
  - !ruby/object:Gem::Version
53
- version: 5.3.0
67
+ version: 5.4.0
54
68
  description: AltText helps with accessibility by generating alt text for images.
55
69
  email:
56
70
  - ajk5603@psu.edu
@@ -61,6 +75,7 @@ files:
61
75
  - LICENSE.txt
62
76
  - README.md
63
77
  - bin/alt_text
78
+ - bin/check_dependencies.rb
64
79
  - lib/alt_text.rb
65
80
  - lib/alt_text/client.rb
66
81
  - lib/alt_text/llm_registry.rb
@@ -85,7 +100,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
85
100
  - !ruby/object:Gem::Version
86
101
  version: '0'
87
102
  requirements: []
88
- rubygems_version: 3.6.2
103
+ rubygems_version: 3.6.9
89
104
  specification_version: 4
90
105
  summary: Generates alt text
91
106
  test_files: []