huginn_ollama_agent 0.1.10 → 0.1.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: '079d5ac6979bc3f61c17450391c7fba830471e7db33a575fdc92514b4968b5c2'
-  data.tar.gz: f3c36631d72eb39c5f980cac4b25c1a7053f9440fd363c8e5f3b3f8d30019f44
+  metadata.gz: f2528d73ee7c66815ff599fea1ca57382036a367670849a0c9d1180c94dd4035
+  data.tar.gz: 73f49b9773db8c5015da9d42c01764023c61d76bfa9badac00b289446c20ca8f
 SHA512:
-  metadata.gz: b0b0bc0dfae1f61aa4e07b1969253aab68d36b117be49e9a09da1af7e5274c081c72b41ef09f11c46754b3a5206c6478850ffb627b0e0af7a697d3b5e4f62d5d
-  data.tar.gz: 6ef17683e955c276c52f310886b06862b2906168518a2040b7bfae3b830f4c59597f814838af1ffd52cef96ea777d4bf35c9e694b2b57597a1ed3a27174b009a
+  metadata.gz: 5ad7f7f1881cb615762106c7e0f1d25aefe031515c61d1fc3c1e498ac130e8c84b3d449cb12444823c0ba0fb7154208c1255c563c09606a73e4d0bc167c931b1
+  data.tar.gz: 853c508e1a36aa8c0b1f96d73703c05fd73a220097f374080bd87bf2c99aa2cd004602ffea18c6cc485c7c3c7a3517079a0beec49f8ae98c800421e4597a9543

data/lib/huginn_ollama_agent/ollama_agent.rb

@@ -17,7 +17,7 @@ module Agents
 
       `prompt` the prompt to generate a response.
 
-      `image` is a base64-encoded image (for multimodal models such as llava).
+      `image` is a base64-encoded image or URL (for multimodal models such as llava).
 
       `context` the context parameter returned from a previous request to /generate, this can be used to keep a short conversational memory.
 
@@ -249,6 +249,38 @@ module Agents
       false
     end
 
+    def encode_to_base64(data)
+      return Base64.strict_encode64(data)
+    end
+
+    def download_and_convert_to_base64(url)
+      begin
+        uri = URI(url)
+        response = Net::HTTP.get_response(uri)
+
+        log_curl_output(response.code,response.body)
+
+        if response.is_a?(Net::HTTPSuccess)
+          file_data = response.body
+          base64_data = encode_to_base64(file_data)
+          return base64_data
+        end
+      end
+    end
+
+    def detect_image_source(input)
+      if input =~ URI::DEFAULT_PARSER.regexp[:ABS_URI]
+        download_and_convert_to_base64(interpolated['image'])
+      elsif input.match?(/\A(?:[A-Za-z0-9+\/]{4})*(?:[A-Za-z0-9+\/]{2}==|[A-Za-z0-9+\/]{3}=)?\z/)
+        begin
+          decoded = Base64.strict_decode64(input)
+          if Base64.strict_encode64(decoded) == input
+            return input
+          end
+        end
+      end
+    end
+
     def generate_completion()
 
       request_payload = {}
@@ -257,7 +289,7 @@ module Agents
       request_payload['stream'] = boolify(interpolated['stream'])
       request_payload['raw'] = boolify(interpolated['raw'])
       request_payload['context'] = context if !interpolated['context'].empty?
-      request_payload['images'] = ["#{interpolated['image']}"] if !interpolated['image'].empty?
+      request_payload['images'] = ["#{detect_image_source(interpolated['image'])}"] if !interpolated['image'].empty?
 
       if check_remote_model()
         if interpolated['debug'] == 'true'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: huginn_ollama_agent
 version: !ruby/object:Gem::Version
-  version: 0.1.10
+  version: 0.1.11
 platform: ruby
 authors:
 - Nicolas Germain
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2024-03-27 00:00:00.000000000 Z
+date: 2024-03-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler