raif 1.4.0 → 1.5.0
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.
- checksums.yaml +4 -4
- data/README.md +2 -2
- data/app/assets/builds/raif_admin.css +40 -2
- data/app/assets/builds/raif_admin_sprockets.js +2709 -0
- data/app/assets/javascript/raif/admin/copy_to_clipboard_controller.js +132 -0
- data/app/assets/javascript/raif/admin/cost_estimate_controller.js +80 -0
- data/app/assets/javascript/raif/admin/judge_config_controller.js +23 -0
- data/app/assets/javascript/raif/admin/select_all_checkboxes_controller.js +33 -0
- data/app/assets/javascript/raif/admin/sortable_table_controller.js +51 -0
- data/app/assets/javascript/raif/admin/table_search_controller.js +15 -0
- data/app/assets/javascript/raif/admin/tom_select_controller.js +33 -0
- data/app/assets/javascript/raif_admin.js +23 -0
- data/app/assets/javascript/raif_admin_sprockets.js +24 -0
- data/app/assets/stylesheets/raif_admin.scss +50 -1
- data/app/controllers/raif/admin/agents_controller.rb +27 -1
- data/app/controllers/raif/admin/configs_controller.rb +1 -0
- data/app/controllers/raif/admin/llms_controller.rb +27 -0
- data/app/controllers/raif/admin/model_completions_controller.rb +6 -0
- data/app/controllers/raif/admin/prompt_studio/agents_controller.rb +25 -0
- data/app/controllers/raif/admin/prompt_studio/base_controller.rb +32 -0
- data/app/controllers/raif/admin/prompt_studio/batch_runs_controller.rb +102 -0
- data/app/controllers/raif/admin/prompt_studio/conversations_controller.rb +25 -0
- data/app/controllers/raif/admin/prompt_studio/tasks_controller.rb +64 -0
- data/app/controllers/raif/admin/tasks_controller.rb +5 -0
- data/app/helpers/raif/application_helper.rb +40 -0
- data/app/jobs/raif/prompt_studio_batch_run_item_job.rb +11 -0
- data/app/jobs/raif/prompt_studio_batch_run_job.rb +15 -0
- data/app/jobs/raif/prompt_studio_task_run_job.rb +36 -0
- data/app/models/raif/agent.rb +36 -5
- data/app/models/raif/agents/native_tool_calling_agent.rb +101 -19
- data/app/models/raif/concerns/has_prompt_templates.rb +88 -0
- data/app/models/raif/concerns/has_runtime_duration.rb +41 -0
- data/app/models/raif/concerns/json_schema_definition.rb +16 -3
- data/app/models/raif/concerns/llm_prompt_caching.rb +20 -0
- data/app/models/raif/concerns/llms/anthropic/message_formatting.rb +6 -0
- data/app/models/raif/concerns/llms/anthropic/tool_formatting.rb +5 -1
- data/app/models/raif/concerns/llms/bedrock/message_formatting.rb +7 -0
- data/app/models/raif/concerns/llms/bedrock/tool_formatting.rb +4 -0
- data/app/models/raif/concerns/llms/google/message_formatting.rb +5 -2
- data/app/models/raif/concerns/llms/google/tool_formatting.rb +4 -0
- data/app/models/raif/concerns/llms/message_formatting.rb +30 -0
- data/app/models/raif/concerns/llms/open_ai_completions/response_tool_calls.rb +1 -1
- data/app/models/raif/concerns/llms/open_ai_completions/tool_formatting.rb +4 -0
- data/app/models/raif/concerns/llms/open_ai_responses/tool_formatting.rb +4 -0
- data/app/models/raif/concerns/provider_managed_tool_calls.rb +162 -0
- data/app/models/raif/conversation.rb +24 -3
- data/app/models/raif/conversation_entry.rb +6 -3
- data/app/models/raif/embedding_models/bedrock.rb +10 -1
- data/app/models/raif/embedding_models/google.rb +37 -0
- data/app/models/raif/evals/llm_judge.rb +70 -0
- data/{lib → app/models}/raif/evals/llm_judges/binary.rb +38 -0
- data/{lib → app/models}/raif/evals/llm_judges/comparative.rb +38 -0
- data/{lib → app/models}/raif/evals/llm_judges/scored.rb +38 -0
- data/{lib → app/models}/raif/evals/llm_judges/summarization.rb +38 -0
- data/app/models/raif/llm.rb +82 -7
- data/app/models/raif/llms/anthropic.rb +26 -4
- data/app/models/raif/llms/bedrock.rb +59 -5
- data/app/models/raif/llms/google.rb +28 -2
- data/app/models/raif/llms/open_ai_base.rb +4 -0
- data/app/models/raif/llms/open_ai_completions.rb +9 -2
- data/app/models/raif/llms/open_ai_responses.rb +9 -2
- data/app/models/raif/llms/open_router.rb +10 -3
- data/app/models/raif/model_completion.rb +75 -34
- data/app/models/raif/model_tool.rb +45 -3
- data/app/models/raif/model_tool_invocation.rb +31 -1
- data/app/models/raif/prompt_studio_batch_run.rb +155 -0
- data/app/models/raif/prompt_studio_batch_run_item.rb +220 -0
- data/app/models/raif/streaming_responses/bedrock.rb +60 -1
- data/app/models/raif/task.rb +30 -6
- data/app/views/layouts/raif/admin.html.erb +31 -1
- data/app/views/raif/admin/agents/_agent.html.erb +1 -0
- data/app/views/raif/admin/agents/index.html.erb +48 -0
- data/app/views/raif/admin/agents/show.html.erb +4 -0
- data/app/views/raif/admin/llms/index.html.erb +110 -0
- data/app/views/raif/admin/model_completions/_model_completion.html.erb +3 -7
- data/app/views/raif/admin/model_completions/index.html.erb +14 -1
- data/app/views/raif/admin/model_completions/show.html.erb +164 -55
- data/app/views/raif/admin/model_tool_invocations/index.html.erb +1 -1
- data/app/views/raif/admin/model_tool_invocations/show.html.erb +18 -0
- data/app/views/raif/admin/prompt_studio/agents/index.html.erb +56 -0
- data/app/views/raif/admin/prompt_studio/agents/show.html.erb +57 -0
- data/app/views/raif/admin/prompt_studio/batch_runs/_batch_run_item.html.erb +54 -0
- data/app/views/raif/admin/prompt_studio/batch_runs/_judge_config_fields.html.erb +76 -0
- data/app/views/raif/admin/prompt_studio/batch_runs/_judge_detail_modal.html.erb +27 -0
- data/app/views/raif/admin/prompt_studio/batch_runs/_modal.html.erb +35 -0
- data/app/views/raif/admin/prompt_studio/batch_runs/_progress.html.erb +78 -0
- data/app/views/raif/admin/prompt_studio/batch_runs/show.html.erb +49 -0
- data/app/views/raif/admin/prompt_studio/conversations/index.html.erb +48 -0
- data/app/views/raif/admin/prompt_studio/conversations/show.html.erb +36 -0
- data/app/views/raif/admin/prompt_studio/shared/_nav_tabs.html.erb +17 -0
- data/app/views/raif/admin/prompt_studio/shared/_prompt_comparison.html.erb +87 -0
- data/app/views/raif/admin/prompt_studio/shared/_type_filter.html.erb +54 -0
- data/app/views/raif/admin/prompt_studio/tasks/_task_result.html.erb +145 -0
- data/app/views/raif/admin/prompt_studio/tasks/_task_row.html.erb +12 -0
- data/app/views/raif/admin/prompt_studio/tasks/_task_type_filter.html.erb +58 -0
- data/app/views/raif/admin/prompt_studio/tasks/_tasks_table.html.erb +22 -0
- data/app/views/raif/admin/prompt_studio/tasks/index.html.erb +35 -0
- data/app/views/raif/admin/prompt_studio/tasks/show.html.erb +19 -0
- data/app/views/raif/admin/tasks/_task.html.erb +1 -0
- data/app/views/raif/admin/tasks/index.html.erb +17 -5
- data/app/views/raif/admin/tasks/show.html.erb +20 -0
- data/app/views/raif/conversation_entries/_message.html.erb +10 -6
- data/config/importmap.rb +8 -0
- data/config/locales/admin.en.yml +128 -0
- data/config/locales/en.yml +36 -2
- data/config/routes.rb +8 -0
- data/db/migrate/20260307000000_add_prompt_studio_run_to_raif_tasks.rb +7 -0
- data/db/migrate/20260308000000_create_raif_prompt_studio_batch_runs.rb +27 -0
- data/db/migrate/20260308000001_create_raif_prompt_studio_batch_run_items.rb +24 -0
- data/db/migrate/20260407000000_add_cache_token_columns_to_raif_model_completions.rb +8 -0
- data/lib/generators/raif/agent/agent_generator.rb +18 -0
- data/lib/generators/raif/agent/templates/agent.rb.tt +7 -5
- data/lib/generators/raif/agent/templates/system_prompt.erb.tt +3 -0
- data/lib/generators/raif/conversation/conversation_generator.rb +19 -1
- data/lib/generators/raif/conversation/templates/system_prompt.erb.tt +4 -0
- data/lib/generators/raif/install/templates/initializer.rb +68 -27
- data/lib/generators/raif/task/task_generator.rb +18 -0
- data/lib/generators/raif/task/templates/prompt.erb.tt +4 -0
- data/lib/generators/raif/task/templates/task.rb.tt +9 -8
- data/lib/raif/configuration.rb +10 -0
- data/lib/raif/embedding_model_registry.rb +8 -0
- data/lib/raif/engine.rb +16 -1
- data/lib/raif/errors/blank_response_error.rb +8 -0
- data/lib/raif/errors/prompt_template_error.rb +15 -0
- data/lib/raif/errors.rb +2 -0
- data/lib/raif/evals.rb +0 -6
- data/lib/raif/llm_registry.rb +230 -9
- data/lib/raif/prompt_studio_comparison_builder.rb +138 -0
- data/lib/raif/token_estimator.rb +28 -0
- data/lib/raif/version.rb +1 -1
- data/lib/raif.rb +2 -0
- data/spec/support/rspec_helpers.rb +7 -1
- data/spec/support/test_task.rb +9 -0
- data/spec/support/test_template_task.rb +41 -0
- metadata +65 -7
- data/lib/raif/evals/llm_judge.rb +0 -32
- /data/{lib → app/models}/raif/evals/scoring_rubric.rb +0 -0
|
@@ -58,6 +58,10 @@
|
|
|
58
58
|
<div class="col-5"><strong><%= t("raif.admin.common.completed_at") %>:</strong></div>
|
|
59
59
|
<div class="col-7"><small><%= @model_completion.completed_at&.rfc822 || "-" %></small></div>
|
|
60
60
|
</div>
|
|
61
|
+
<div class="row mb-2">
|
|
62
|
+
<div class="col-5"><strong><%= t("raif.admin.common.duration") %>:</strong></div>
|
|
63
|
+
<div class="col-7"><small><%= @model_completion.runtime_duration %></small></div>
|
|
64
|
+
</div>
|
|
61
65
|
<div class="row mb-2">
|
|
62
66
|
<div class="col-5"><strong><%= t("raif.admin.common.prompt_tokens") %>:</strong></div>
|
|
63
67
|
<div class="col-7">
|
|
@@ -97,71 +101,88 @@
|
|
|
97
101
|
<% end %>
|
|
98
102
|
</div>
|
|
99
103
|
</div>
|
|
104
|
+
<hr class="my-2">
|
|
105
|
+
<div class="row mb-2">
|
|
106
|
+
<div class="col-5"><strong><%= t("raif.admin.common.cache_read_input_tokens") %>:</strong></div>
|
|
107
|
+
<div class="col-7"><%= @model_completion.cache_read_input_tokens ? number_with_delimiter(@model_completion.cache_read_input_tokens) : "-" %></div>
|
|
108
|
+
</div>
|
|
109
|
+
<div class="row mb-2">
|
|
110
|
+
<div class="col-5"><strong><%= t("raif.admin.common.cache_creation_input_tokens") %>:</strong></div>
|
|
111
|
+
<div class="col-7"><%= @model_completion.cache_creation_input_tokens ? number_with_delimiter(@model_completion.cache_creation_input_tokens) : "-" %></div>
|
|
112
|
+
</div>
|
|
100
113
|
</div>
|
|
101
114
|
</div>
|
|
102
115
|
<% if @model_completion.failed? %>
|
|
103
|
-
|
|
104
|
-
|
|
105
|
-
|
|
106
|
-
|
|
107
|
-
|
|
108
|
-
|
|
109
|
-
|
|
110
|
-
|
|
111
|
-
|
|
112
|
-
|
|
113
|
-
|
|
114
|
-
|
|
115
|
-
|
|
116
|
+
<hr>
|
|
117
|
+
<div class="row mb-2">
|
|
118
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.failed_at") %>:</strong></div>
|
|
119
|
+
<div class="col-md-10"><small><%= @model_completion.failed_at&.rfc822 || "-" %></small></div>
|
|
120
|
+
</div>
|
|
121
|
+
<div class="row mb-2">
|
|
122
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.failure_error") %>:</strong></div>
|
|
123
|
+
<div class="col-md-10"><code><%= @model_completion.failure_error %></code></div>
|
|
124
|
+
</div>
|
|
125
|
+
<div class="row mb-2">
|
|
126
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.failure_reason") %>:</strong></div>
|
|
127
|
+
<div class="col-md-10"><%= @model_completion.failure_reason %></div>
|
|
128
|
+
</div>
|
|
116
129
|
<% end %>
|
|
117
130
|
</div>
|
|
118
131
|
</div>
|
|
119
132
|
|
|
120
133
|
<% if @model_completion.messages.present? %>
|
|
121
|
-
<div class="card mb-4">
|
|
122
|
-
|
|
123
|
-
|
|
124
|
-
|
|
125
|
-
|
|
126
|
-
|
|
127
|
-
|
|
128
|
-
|
|
129
|
-
|
|
130
|
-
|
|
131
|
-
|
|
132
|
-
|
|
133
|
-
|
|
134
|
+
<div class="card mb-4">
|
|
135
|
+
<div class="card-header">
|
|
136
|
+
<h5 class="mb-0"><%= t("raif.admin.common.messages") %></h5>
|
|
137
|
+
</div>
|
|
138
|
+
<div class="card-body">
|
|
139
|
+
<% @model_completion.messages.each do |message| %>
|
|
140
|
+
<div class="mb-3">
|
|
141
|
+
<% if message["type"] == "function_call" %>
|
|
142
|
+
<strong><%= t("raif.admin.common.tool_call") %>:</strong>
|
|
143
|
+
<div class="border rounded p-2 mt-2 bg-light">
|
|
144
|
+
<div><strong><%= t("raif.admin.common.tool_name") %>:</strong> <%= message["name"] %></div>
|
|
145
|
+
<div><strong><%= t("raif.admin.common.arguments") %>:</strong></div>
|
|
146
|
+
<pre class="mb-0 mt-1"><%= pretty_json(message["arguments"]) %></pre>
|
|
147
|
+
</div>
|
|
148
|
+
<% elsif message["type"] == "function_call_output" %>
|
|
149
|
+
<strong><%= t("raif.admin.common.tool_result") %>:</strong>
|
|
150
|
+
<pre class="mt-2"><%= pretty_json(message["output"]) %></pre>
|
|
151
|
+
<% else %>
|
|
152
|
+
<% if message["role"].present? %>
|
|
153
|
+
<strong><%= message["role"].titleize %>:</strong>
|
|
154
|
+
<% end %>
|
|
155
|
+
<% if message["content"].present? %>
|
|
156
|
+
<pre class="mt-2"><%= message["content"] %></pre>
|
|
157
|
+
<% end %>
|
|
134
158
|
|
|
135
|
-
|
|
136
|
-
|
|
137
|
-
|
|
138
|
-
|
|
139
|
-
|
|
140
|
-
|
|
141
|
-
|
|
142
|
-
|
|
143
|
-
|
|
144
|
-
tool_call.dig("function", "arguments")
|
|
145
|
-
end %></pre>
|
|
159
|
+
<% if message["tool_calls"].present? %>
|
|
160
|
+
<div class="mt-2">
|
|
161
|
+
<% message["tool_calls"].each do |tool_call| %>
|
|
162
|
+
<div class="border rounded p-2 mb-2 bg-light">
|
|
163
|
+
<div><strong><%= t("raif.admin.common.tool_name") %>:</strong> <%= tool_call.dig("function", "name") %></div>
|
|
164
|
+
<div><strong><%= t("raif.admin.common.arguments") %>:</strong></div>
|
|
165
|
+
<pre class="mb-0 mt-1"><%= pretty_json(tool_call.dig("function", "arguments")) %></pre>
|
|
166
|
+
</div>
|
|
167
|
+
<% end %>
|
|
146
168
|
</div>
|
|
147
169
|
<% end %>
|
|
148
|
-
|
|
149
|
-
|
|
150
|
-
|
|
151
|
-
|
|
170
|
+
<% end %>
|
|
171
|
+
</div>
|
|
172
|
+
<% end %>
|
|
173
|
+
</div>
|
|
152
174
|
</div>
|
|
153
|
-
</div>
|
|
154
175
|
<% end %>
|
|
155
176
|
|
|
156
177
|
<% if @model_completion.system_prompt.present? %>
|
|
157
|
-
<div class="card mb-4">
|
|
158
|
-
|
|
159
|
-
|
|
160
|
-
|
|
161
|
-
|
|
162
|
-
|
|
178
|
+
<div class="card mb-4">
|
|
179
|
+
<div class="card-header">
|
|
180
|
+
<h5 class="mb-0"><%= t("raif.admin.common.system_prompt") %></h5>
|
|
181
|
+
</div>
|
|
182
|
+
<div class="card-body">
|
|
183
|
+
<pre class="pre-wrap"><%= @model_completion.system_prompt %></pre>
|
|
184
|
+
</div>
|
|
163
185
|
</div>
|
|
164
|
-
</div>
|
|
165
186
|
<% end %>
|
|
166
187
|
|
|
167
188
|
<div class="card mb-4">
|
|
@@ -176,7 +197,7 @@
|
|
|
176
197
|
<h5 class="mb-0 mt-4"><%= t("raif.admin.common.prettified") %></h5>
|
|
177
198
|
<% begin %>
|
|
178
199
|
<pre class="pre-wrap"><%= JSON.pretty_generate(@model_completion.parsed_response) %></pre>
|
|
179
|
-
|
|
200
|
+
<% rescue StandardError %>
|
|
180
201
|
<%= t("raif.admin.common.invalid_json") %>
|
|
181
202
|
<% end %>
|
|
182
203
|
<% elsif @model_completion.response_format_html? %>
|
|
@@ -206,22 +227,95 @@
|
|
|
206
227
|
</div>
|
|
207
228
|
</div>
|
|
208
229
|
|
|
230
|
+
<% if @model_completion.provider_managed_tool_calls.present? %>
|
|
231
|
+
<div class="card mb-4">
|
|
232
|
+
<div class="card-header">
|
|
233
|
+
<h5 class="mb-0"><%= t("raif.admin.common.provider_managed_tool_calls") %></h5>
|
|
234
|
+
</div>
|
|
235
|
+
<div class="card-body">
|
|
236
|
+
<% @model_completion.provider_managed_tool_calls.each do |tool_call| %>
|
|
237
|
+
<div class="border rounded p-3 mb-3 bg-light">
|
|
238
|
+
<div class="d-flex justify-content-between align-items-start gap-3">
|
|
239
|
+
<div>
|
|
240
|
+
<div><strong><%= t("raif.admin.common.tool_name") %>:</strong> <%= tool_call["tool_name"].to_s.titleize %></div>
|
|
241
|
+
<% if tool_call["provider_tool_call_id"].present? %>
|
|
242
|
+
<div><strong><%= t("raif.admin.common.provider_tool_call_id") %>:</strong> <code><%= tool_call["provider_tool_call_id"] %></code></div>
|
|
243
|
+
<% end %>
|
|
244
|
+
<% if tool_call["inferred"] %>
|
|
245
|
+
<div><small class="text-muted"><%= t("raif.admin.common.inferred_from_citations") %></small></div>
|
|
246
|
+
<% end %>
|
|
247
|
+
</div>
|
|
248
|
+
<% if tool_call["status"].present? %>
|
|
249
|
+
<span class="badge bg-info text-dark"><%= tool_call["status"].to_s.titleize %></span>
|
|
250
|
+
<% end %>
|
|
251
|
+
</div>
|
|
252
|
+
|
|
253
|
+
<% if tool_call["arguments"].present? %>
|
|
254
|
+
<div class="mt-3"><strong><%= t("raif.admin.common.arguments") %>:</strong></div>
|
|
255
|
+
<pre class="mb-0 mt-1"><%= begin
|
|
256
|
+
JSON.pretty_generate(tool_call["arguments"])
|
|
257
|
+
rescue StandardError
|
|
258
|
+
tool_call["arguments"]
|
|
259
|
+
end %></pre>
|
|
260
|
+
<% end %>
|
|
261
|
+
|
|
262
|
+
<% if tool_call["sources"].present? %>
|
|
263
|
+
<div class="mt-3">
|
|
264
|
+
<strong><%= t("raif.admin.common.sources") %>:</strong>
|
|
265
|
+
<div class="mt-2">
|
|
266
|
+
<% tool_call["sources"].each do |source| %>
|
|
267
|
+
<div class="border rounded p-2 mb-2 bg-white">
|
|
268
|
+
<% if source["url"].present? %>
|
|
269
|
+
<div>
|
|
270
|
+
<a href="<%= source["url"] %>" target="_blank" rel="noopener" class="text-decoration-none">
|
|
271
|
+
<%= source["title"].presence || source["url"] %>
|
|
272
|
+
</a>
|
|
273
|
+
</div>
|
|
274
|
+
<small class="text-muted d-block"><%= source["url"] %></small>
|
|
275
|
+
<% else %>
|
|
276
|
+
<div><%= source["title"] %></div>
|
|
277
|
+
<% end %>
|
|
278
|
+
<% if source["page_age"].present? %>
|
|
279
|
+
<small class="text-muted d-block"><%= source["page_age"] %></small>
|
|
280
|
+
<% end %>
|
|
281
|
+
</div>
|
|
282
|
+
<% end %>
|
|
283
|
+
</div>
|
|
284
|
+
</div>
|
|
285
|
+
<% elsif tool_call["raw_result"].present? %>
|
|
286
|
+
<div class="mt-3"><strong><%= t("raif.admin.common.result") %>:</strong></div>
|
|
287
|
+
<pre class="mb-0 mt-1"><%= begin
|
|
288
|
+
JSON.pretty_generate(tool_call["raw_result"])
|
|
289
|
+
rescue StandardError
|
|
290
|
+
tool_call["raw_result"]
|
|
291
|
+
end %></pre>
|
|
292
|
+
<% end %>
|
|
293
|
+
</div>
|
|
294
|
+
<% end %>
|
|
295
|
+
</div>
|
|
296
|
+
</div>
|
|
297
|
+
<% end %>
|
|
298
|
+
|
|
209
299
|
<div class="card mb-4">
|
|
210
300
|
<div class="card-header">
|
|
211
301
|
<h5 class="mb-0"><%= t("raif.admin.common.citations") %></h5>
|
|
212
302
|
</div>
|
|
213
303
|
<div class="card-body">
|
|
214
|
-
<% if @model_completion.
|
|
215
|
-
<% @model_completion.
|
|
304
|
+
<% if @model_completion.sanitized_citations.present? %>
|
|
305
|
+
<% @model_completion.sanitized_citations.each_with_index do |citation, index| %>
|
|
216
306
|
<div class="mb-3 p-3 border rounded">
|
|
217
307
|
<div class="d-flex align-items-start">
|
|
218
308
|
<span class="badge bg-primary me-3"><%= index + 1 %></span>
|
|
219
309
|
<div class="flex-grow-1">
|
|
220
310
|
<h6 class="mb-1">
|
|
221
|
-
|
|
311
|
+
<% if citation["url"].present? %>
|
|
312
|
+
<a href="<%= citation["url"] %>" target="_blank" rel="noopener" class="text-decoration-none">
|
|
313
|
+
<%= citation["title"] %>
|
|
314
|
+
<i class="bi bi-box-arrow-up-right ms-1" style="font-size: 0.8em;"></i>
|
|
315
|
+
</a>
|
|
316
|
+
<% else %>
|
|
222
317
|
<%= citation["title"] %>
|
|
223
|
-
|
|
224
|
-
</a>
|
|
318
|
+
<% end %>
|
|
225
319
|
</h6>
|
|
226
320
|
<small class="text-muted"><%= citation["url"] %></small>
|
|
227
321
|
</div>
|
|
@@ -233,3 +327,18 @@
|
|
|
233
327
|
<% end %>
|
|
234
328
|
</div>
|
|
235
329
|
</div>
|
|
330
|
+
|
|
331
|
+
<% if @model_completion.response_array.present? %>
|
|
332
|
+
<div class="card mb-4">
|
|
333
|
+
<div class="card-header">
|
|
334
|
+
<h5 class="mb-0"><%= t("raif.admin.common.response_array") %></h5>
|
|
335
|
+
</div>
|
|
336
|
+
<div class="card-body">
|
|
337
|
+
<pre class="pre-wrap"><%= begin
|
|
338
|
+
JSON.pretty_generate(@model_completion.response_array)
|
|
339
|
+
rescue StandardError
|
|
340
|
+
@model_completion.response_array
|
|
341
|
+
end %></pre>
|
|
342
|
+
</div>
|
|
343
|
+
</div>
|
|
344
|
+
<% end %>
|
|
@@ -14,7 +14,7 @@
|
|
|
14
14
|
<label for="tool_types"><%= t("raif.admin.common.tool_type") %></label>
|
|
15
15
|
<%= select_tag :tool_types,
|
|
16
16
|
options_for_select([[t("raif.admin.common.all"), "all"]] + @tool_types.map{|type| [type, type] }, @selected_type),
|
|
17
|
-
{ class: "form-select" } %>
|
|
17
|
+
{ class: "form-select", data: { controller: "raif--tom-select" } } %>
|
|
18
18
|
</div>
|
|
19
19
|
</div>
|
|
20
20
|
<div class="col-md-2">
|
|
@@ -64,3 +64,21 @@
|
|
|
64
64
|
</div>
|
|
65
65
|
</div>
|
|
66
66
|
<% end %>
|
|
67
|
+
|
|
68
|
+
<% if @model_tool_invocation.admin_observation_available? %>
|
|
69
|
+
<div class="card mb-4">
|
|
70
|
+
<div class="card-header">
|
|
71
|
+
<h5 class="mb-0"><%= t(".observation_sent_to_model") %></h5>
|
|
72
|
+
</div>
|
|
73
|
+
<% if @model_tool_invocation.admin_observation_error.present? %>
|
|
74
|
+
<div class="card-body text-muted">
|
|
75
|
+
<%= t(".observation_unavailable", error: @model_tool_invocation.admin_observation_error) %>
|
|
76
|
+
</div>
|
|
77
|
+
<% else %>
|
|
78
|
+
<div class="card-body">
|
|
79
|
+
<p class="text-muted small"><%= t(".observation_disclaimer") %></p>
|
|
80
|
+
<pre class="pre-wrap"><%= @model_tool_invocation.admin_observation %></pre>
|
|
81
|
+
</div>
|
|
82
|
+
<% end %>
|
|
83
|
+
</div>
|
|
84
|
+
<% end %>
|
|
@@ -0,0 +1,56 @@
|
|
|
1
|
+
<div class="d-flex justify-content-between align-items-center my-4">
|
|
2
|
+
<h1 class="mb-0"><%= t("raif.admin.prompt_studio.common.prompt_studio") %></h1>
|
|
3
|
+
</div>
|
|
4
|
+
|
|
5
|
+
<%= render "raif/admin/prompt_studio/shared/nav_tabs", active_tab: :agents %>
|
|
6
|
+
|
|
7
|
+
<%= render "raif/admin/prompt_studio/shared/type_filter",
|
|
8
|
+
filter_path: raif.admin_prompt_studio_agents_path,
|
|
9
|
+
param_name: :agent_type,
|
|
10
|
+
types: @agent_types,
|
|
11
|
+
selected_type: @selected_type,
|
|
12
|
+
llm_model_keys: @llm_model_keys %>
|
|
13
|
+
|
|
14
|
+
<% if @selected_type.present? && @agents.present? %>
|
|
15
|
+
<div class="table-responsive">
|
|
16
|
+
<table class="table table-striped">
|
|
17
|
+
<thead>
|
|
18
|
+
<tr>
|
|
19
|
+
<th><%= t("raif.admin.common.id") %></th>
|
|
20
|
+
<th><%= t("raif.admin.common.type") %></th>
|
|
21
|
+
<th><%= t("raif.admin.common.created_at") %></th>
|
|
22
|
+
<th><%= t("raif.admin.common.model") %></th>
|
|
23
|
+
<th><%= t("raif.admin.common.status") %></th>
|
|
24
|
+
<th></th>
|
|
25
|
+
</tr>
|
|
26
|
+
</thead>
|
|
27
|
+
<tbody>
|
|
28
|
+
<% @agents.each do |agent| %>
|
|
29
|
+
<tr>
|
|
30
|
+
<td><%= agent.id %></td>
|
|
31
|
+
<td><%= agent.type %></td>
|
|
32
|
+
<td><%= agent.created_at.rfc822 %></td>
|
|
33
|
+
<td><%= agent.llm_model_key %></td>
|
|
34
|
+
<td>
|
|
35
|
+
<% if agent.completed_at? %>
|
|
36
|
+
<span class="badge bg-success"><%= t("raif.admin.common.completed") %></span>
|
|
37
|
+
<% elsif agent.failed_at? %>
|
|
38
|
+
<span class="badge bg-danger"><%= t("raif.admin.common.failed") %></span>
|
|
39
|
+
<% elsif agent.started_at? %>
|
|
40
|
+
<span class="badge bg-warning text-dark"><%= t("raif.admin.common.in_progress") %></span>
|
|
41
|
+
<% end %>
|
|
42
|
+
</td>
|
|
43
|
+
<td>
|
|
44
|
+
<%= link_to "View", raif.admin_prompt_studio_agent_path(agent), class: "btn btn-sm btn-outline-primary" %>
|
|
45
|
+
</td>
|
|
46
|
+
</tr>
|
|
47
|
+
<% end %>
|
|
48
|
+
</tbody>
|
|
49
|
+
</table>
|
|
50
|
+
</div>
|
|
51
|
+
<div class="mt-3"><%== pagy_bootstrap_nav(@pagy) if @pagy.pages > 1 %></div>
|
|
52
|
+
<% elsif @selected_type.present? %>
|
|
53
|
+
<div class="alert alert-info">
|
|
54
|
+
<%= t("raif.admin.prompt_studio.common.no_instances") %>
|
|
55
|
+
</div>
|
|
56
|
+
<% end %>
|
|
@@ -0,0 +1,57 @@
|
|
|
1
|
+
<div class="d-flex justify-content-between align-items-center my-4">
|
|
2
|
+
<h1 class="mb-0"><%= t("raif.admin.prompt_studio.agents.show.page_title", id: @agent.id) %></h1>
|
|
3
|
+
<%= link_to t("raif.admin.prompt_studio.common.back"),
|
|
4
|
+
raif.admin_prompt_studio_agents_path(agent_type: @agent.type),
|
|
5
|
+
class: "btn btn-outline-secondary" %>
|
|
6
|
+
</div>
|
|
7
|
+
|
|
8
|
+
<div class="card mb-4">
|
|
9
|
+
<div class="card-header">
|
|
10
|
+
<h5 class="mb-0"><%= t("raif.admin.prompt_studio.agents.show.agent_details") %></h5>
|
|
11
|
+
</div>
|
|
12
|
+
<div class="card-body">
|
|
13
|
+
<div class="row mb-3">
|
|
14
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.id") %>:</strong></div>
|
|
15
|
+
<div class="col-md-10"><%= @agent.id %></div>
|
|
16
|
+
</div>
|
|
17
|
+
<div class="row mb-3">
|
|
18
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.type") %>:</strong></div>
|
|
19
|
+
<div class="col-md-10"><%= @agent.type %></div>
|
|
20
|
+
</div>
|
|
21
|
+
<div class="row mb-3">
|
|
22
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.model") %>:</strong></div>
|
|
23
|
+
<div class="col-md-10"><%= @agent.llm_model_key %></div>
|
|
24
|
+
</div>
|
|
25
|
+
<div class="row mb-3">
|
|
26
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.created_at") %>:</strong></div>
|
|
27
|
+
<div class="col-md-10"><%= @agent.created_at.rfc822 %></div>
|
|
28
|
+
</div>
|
|
29
|
+
<% if @agent.task.present? %>
|
|
30
|
+
<div class="row mb-3">
|
|
31
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.task") %>:</strong></div>
|
|
32
|
+
<div class="col-md-10"><%= @agent.task %></div>
|
|
33
|
+
</div>
|
|
34
|
+
<% end %>
|
|
35
|
+
<% if @agent.run_with.present? %>
|
|
36
|
+
<div class="row mb-3">
|
|
37
|
+
<div class="col-md-2"><strong><%= t("raif.admin.common.run_with") %>:</strong></div>
|
|
38
|
+
<div class="col-md-10">
|
|
39
|
+
<pre class="pre-wrap mb-0"><%= JSON.pretty_generate(@agent.run_with) %></pre>
|
|
40
|
+
</div>
|
|
41
|
+
</div>
|
|
42
|
+
<% end %>
|
|
43
|
+
</div>
|
|
44
|
+
</div>
|
|
45
|
+
|
|
46
|
+
<%= render "raif/admin/prompt_studio/shared/prompt_comparison", comparison: @comparison %>
|
|
47
|
+
|
|
48
|
+
<% if @agent.final_answer.present? %>
|
|
49
|
+
<div class="card mb-4">
|
|
50
|
+
<div class="card-header">
|
|
51
|
+
<h5 class="mb-0"><%= t("raif.admin.common.final_answer") %></h5>
|
|
52
|
+
</div>
|
|
53
|
+
<div class="card-body">
|
|
54
|
+
<pre class="pre-wrap"><%= @agent.final_answer %></pre>
|
|
55
|
+
</div>
|
|
56
|
+
</div>
|
|
57
|
+
<% end %>
|
|
@@ -0,0 +1,54 @@
|
|
|
1
|
+
<%= tag.tr id: dom_id(item) do %>
|
|
2
|
+
<td>
|
|
3
|
+
<%= link_to "##{item.source_task_id}", raif.admin_prompt_studio_task_path(item.source_task_id), class: "text-decoration-none" %>
|
|
4
|
+
</td>
|
|
5
|
+
<td>
|
|
6
|
+
<% case item.status %>
|
|
7
|
+
<% when "pending" %>
|
|
8
|
+
<span class="badge bg-secondary"><%= t("raif.admin.prompt_studio.batch_runs.status.pending") %></span>
|
|
9
|
+
<% when "running" %>
|
|
10
|
+
<span class="badge bg-warning text-dark">
|
|
11
|
+
<span class="spinner-border spinner-border-sm me-1" role="status" aria-hidden="true"></span>
|
|
12
|
+
<%= t("raif.admin.prompt_studio.batch_runs.status.running") %>
|
|
13
|
+
</span>
|
|
14
|
+
<% when "judging" %>
|
|
15
|
+
<span class="badge bg-info text-dark">
|
|
16
|
+
<span class="spinner-border spinner-border-sm me-1" role="status" aria-hidden="true"></span>
|
|
17
|
+
<%= t("raif.admin.prompt_studio.batch_runs.status.judging") %>
|
|
18
|
+
</span>
|
|
19
|
+
<% when "completed" %>
|
|
20
|
+
<span class="badge bg-success"><%= t("raif.admin.prompt_studio.batch_runs.status.completed") %></span>
|
|
21
|
+
<% when "failed" %>
|
|
22
|
+
<span class="badge bg-danger"><%= t("raif.admin.prompt_studio.batch_runs.status.failed") %></span>
|
|
23
|
+
<% end %>
|
|
24
|
+
</td>
|
|
25
|
+
<td>
|
|
26
|
+
<% if item.source_task&.raw_response.present? %>
|
|
27
|
+
<%= tag.span truncate(item.source_task.raw_response, length: 80), title: item.source_task.raw_response %>
|
|
28
|
+
<% end %>
|
|
29
|
+
</td>
|
|
30
|
+
<td>
|
|
31
|
+
<% if item.result_task&.raw_response.present? %>
|
|
32
|
+
<%= link_to truncate(item.result_task.raw_response, length: 80),
|
|
33
|
+
raif.admin_prompt_studio_task_path(item.result_task),
|
|
34
|
+
class: "text-decoration-none",
|
|
35
|
+
title: item.result_task.raw_response %>
|
|
36
|
+
<% end %>
|
|
37
|
+
</td>
|
|
38
|
+
<% if item.batch_run.has_judge? %>
|
|
39
|
+
<td>
|
|
40
|
+
<% if item.judge_summary.present? %>
|
|
41
|
+
<a href="#" class="text-decoration-none" data-bs-toggle="modal" data-bs-target="#judge-detail-modal-<%= item.id %>">
|
|
42
|
+
<strong><%= item.judge_summary %></strong>
|
|
43
|
+
</a>
|
|
44
|
+
<% end %>
|
|
45
|
+
</td>
|
|
46
|
+
<td>
|
|
47
|
+
<% if item.judge_reasoning.present? %>
|
|
48
|
+
<a href="#" class="text-decoration-none text-body" data-bs-toggle="modal" data-bs-target="#judge-detail-modal-<%= item.id %>">
|
|
49
|
+
<%= truncate(item.judge_reasoning, length: 80) %>
|
|
50
|
+
</a>
|
|
51
|
+
<% end %>
|
|
52
|
+
</td>
|
|
53
|
+
<% end %>
|
|
54
|
+
<% end %>
|
|
@@ -0,0 +1,76 @@
|
|
|
1
|
+
<div id="judge-config-section" data-controller="raif--judge-config">
|
|
2
|
+
<h6 class="mb-2"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.title") %></h6>
|
|
3
|
+
<p class="text-muted small"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.description") %></p>
|
|
4
|
+
|
|
5
|
+
<div class="mb-3">
|
|
6
|
+
<label for="judge_type" class="form-label"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.type_label") %></label>
|
|
7
|
+
<%= select_tag :judge_type,
|
|
8
|
+
options_for_select([
|
|
9
|
+
[t("raif.admin.prompt_studio.batch_runs.judge_config.type_none"), ""],
|
|
10
|
+
[t("raif.admin.prompt_studio.batch_runs.judge_config.type_binary"), "Raif::Evals::LlmJudges::Binary"],
|
|
11
|
+
[t("raif.admin.prompt_studio.batch_runs.judge_config.type_scored"), "Raif::Evals::LlmJudges::Scored"],
|
|
12
|
+
[t("raif.admin.prompt_studio.batch_runs.judge_config.type_comparative"), "Raif::Evals::LlmJudges::Comparative"],
|
|
13
|
+
[t("raif.admin.prompt_studio.batch_runs.judge_config.type_summarization"), "Raif::Evals::LlmJudges::Summarization"]
|
|
14
|
+
]),
|
|
15
|
+
class: "form-select",
|
|
16
|
+
data: { controller: "raif--tom-select", action: "raif--judge-config#toggle raif--cost-estimate#calculate", raif__cost_estimate_target: "judgeTypeSelect" } %>
|
|
17
|
+
</div>
|
|
18
|
+
|
|
19
|
+
<div class="d-none" data-raif--judge-config-target="binary">
|
|
20
|
+
<div class="mb-3">
|
|
21
|
+
<label for="judge_criteria" class="form-label"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.criteria_label") %></label>
|
|
22
|
+
<%= text_area_tag :judge_criteria,
|
|
23
|
+
nil,
|
|
24
|
+
class: "form-control",
|
|
25
|
+
rows: 2,
|
|
26
|
+
placeholder: t("raif.admin.prompt_studio.batch_runs.judge_config.criteria_placeholder") %>
|
|
27
|
+
</div>
|
|
28
|
+
<div class="form-check mb-3">
|
|
29
|
+
<%= check_box_tag :judge_strict_mode, "1", false, class: "form-check-input" %>
|
|
30
|
+
<label for="judge_strict_mode" class="form-check-label"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.strict_mode_label") %></label>
|
|
31
|
+
</div>
|
|
32
|
+
</div>
|
|
33
|
+
|
|
34
|
+
<div class="d-none" data-raif--judge-config-target="scored">
|
|
35
|
+
<div class="mb-3">
|
|
36
|
+
<label for="judge_scoring_rubric" class="form-label"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.rubric_label") %></label>
|
|
37
|
+
<%= select_tag :judge_scoring_rubric,
|
|
38
|
+
options_for_select([
|
|
39
|
+
[t("raif.admin.prompt_studio.batch_runs.judge_config.rubric_accuracy"), "accuracy"],
|
|
40
|
+
[t("raif.admin.prompt_studio.batch_runs.judge_config.rubric_helpfulness"), "helpfulness"],
|
|
41
|
+
[t("raif.admin.prompt_studio.batch_runs.judge_config.rubric_clarity"), "clarity"]
|
|
42
|
+
]),
|
|
43
|
+
class: "form-select" %>
|
|
44
|
+
</div>
|
|
45
|
+
</div>
|
|
46
|
+
|
|
47
|
+
<div class="d-none" data-raif--judge-config-target="comparative">
|
|
48
|
+
<div class="mb-3">
|
|
49
|
+
<label for="judge_comparison_criteria" class="form-label"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.comparison_criteria_label") %></label>
|
|
50
|
+
<%= text_area_tag :judge_comparison_criteria,
|
|
51
|
+
nil,
|
|
52
|
+
class: "form-control",
|
|
53
|
+
rows: 2,
|
|
54
|
+
placeholder: t("raif.admin.prompt_studio.batch_runs.judge_config.comparison_criteria_placeholder") %>
|
|
55
|
+
</div>
|
|
56
|
+
</div>
|
|
57
|
+
|
|
58
|
+
<div class="d-none" data-raif--judge-config-target="summarization">
|
|
59
|
+
<p class="text-muted small"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.summarization_description") %></p>
|
|
60
|
+
</div>
|
|
61
|
+
|
|
62
|
+
<div class="d-none" data-raif--judge-config-target="sharedOptions">
|
|
63
|
+
<div class="form-check mb-3">
|
|
64
|
+
<%= check_box_tag :judge_include_original_prompt_as_context, "1", true, class: "form-check-input" %>
|
|
65
|
+
<label for="judge_include_original_prompt_as_context" class="form-check-label"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.include_original_prompt_as_context_label") %></label>
|
|
66
|
+
<div class="form-text"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.include_original_prompt_as_context_description") %></div>
|
|
67
|
+
</div>
|
|
68
|
+
<div class="mb-3">
|
|
69
|
+
<label for="judge_llm_model_key" class="form-label"><%= t("raif.admin.prompt_studio.batch_runs.judge_config.judge_model_label") %></label>
|
|
70
|
+
<%= select_tag :judge_llm_model_key,
|
|
71
|
+
llm_model_options,
|
|
72
|
+
class: "form-select",
|
|
73
|
+
data: { controller: "raif--tom-select", raif__cost_estimate_target: "judgeModelSelect", action: "raif--cost-estimate#calculate" } %>
|
|
74
|
+
</div>
|
|
75
|
+
</div>
|
|
76
|
+
</div>
|
|
@@ -0,0 +1,27 @@
|
|
|
1
|
+
<div class="modal fade" id="judge-detail-modal-<%= item.id %>" tabindex="-1" aria-hidden="true">
|
|
2
|
+
<div class="modal-dialog modal-lg">
|
|
3
|
+
<div class="modal-content">
|
|
4
|
+
<div class="modal-header">
|
|
5
|
+
<h5 class="modal-title"><%= t("raif.admin.prompt_studio.batch_runs.show.judge_result") %> — <%= t("raif.admin.common.task") %> #<%= item.source_task_id %></h5>
|
|
6
|
+
<button type="button" class="btn-close" data-bs-dismiss="modal" aria-label="Close"></button>
|
|
7
|
+
</div>
|
|
8
|
+
<div class="modal-body">
|
|
9
|
+
<% if item.judge_summary.present? %>
|
|
10
|
+
<div class="mb-3">
|
|
11
|
+
<h6><%= t("raif.admin.prompt_studio.batch_runs.show.judge_result") %></h6>
|
|
12
|
+
<span class="badge bg-primary fs-6"><%= item.judge_summary %></span>
|
|
13
|
+
</div>
|
|
14
|
+
<% end %>
|
|
15
|
+
<% if item.judge_reasoning.present? %>
|
|
16
|
+
<div>
|
|
17
|
+
<h6><%= t("raif.admin.prompt_studio.batch_runs.show.judge_reasoning") %></h6>
|
|
18
|
+
<pre class="pre-wrap border rounded p-3 bg-light"><%= item.judge_reasoning %></pre>
|
|
19
|
+
</div>
|
|
20
|
+
<% end %>
|
|
21
|
+
</div>
|
|
22
|
+
<div class="modal-footer">
|
|
23
|
+
<button type="button" class="btn btn-secondary" data-bs-dismiss="modal"><%= t("raif.admin.prompt_studio.common.back") %></button>
|
|
24
|
+
</div>
|
|
25
|
+
</div>
|
|
26
|
+
</div>
|
|
27
|
+
</div>
|
|
@@ -0,0 +1,35 @@
|
|
|
1
|
+
<div class="modal fade" id="batch-run-modal" tabindex="-1" aria-labelledby="batch-run-modal-label" aria-hidden="true">
|
|
2
|
+
<div class="modal-dialog modal-lg">
|
|
3
|
+
<div class="modal-content">
|
|
4
|
+
<div class="modal-header">
|
|
5
|
+
<h5 class="modal-title" id="batch-run-modal-label"><%= t("raif.admin.prompt_studio.batch_runs.create.title") %></h5>
|
|
6
|
+
<button type="button" class="btn-close" data-bs-dismiss="modal" aria-label="Close"></button>
|
|
7
|
+
</div>
|
|
8
|
+
<div class="modal-body">
|
|
9
|
+
<div class="mb-3">
|
|
10
|
+
<label for="llm_model_key" class="form-label"><%= t("raif.admin.common.model") %></label>
|
|
11
|
+
<%= select_tag :llm_model_key,
|
|
12
|
+
llm_model_options,
|
|
13
|
+
class: "form-select",
|
|
14
|
+
data: { controller: "raif--tom-select", raif__cost_estimate_target: "modelSelect", action: "raif--cost-estimate#calculate" } %>
|
|
15
|
+
</div>
|
|
16
|
+
|
|
17
|
+
<%= render "raif/admin/prompt_studio/batch_runs/judge_config_fields" %>
|
|
18
|
+
|
|
19
|
+
<div class="alert alert-light border mb-0 mt-3 d-none" data-raif--cost-estimate-target="estimate"></div>
|
|
20
|
+
</div>
|
|
21
|
+
<div class="modal-footer d-flex justify-content-between">
|
|
22
|
+
<span class="text-muted">
|
|
23
|
+
<span data-raif--select-all-checkboxes-target="selectedCount">0</span>
|
|
24
|
+
<%= t("raif.admin.prompt_studio.batch_runs.create.tasks_selected") %>
|
|
25
|
+
</span>
|
|
26
|
+
<div class="d-flex gap-2">
|
|
27
|
+
<button type="button" class="btn btn-secondary" data-bs-dismiss="modal"><%= t("raif.admin.prompt_studio.common.back") %></button>
|
|
28
|
+
<%= submit_tag t("raif.admin.prompt_studio.batch_runs.create.submit"),
|
|
29
|
+
class: "btn btn-primary",
|
|
30
|
+
data: { confirm: t("raif.admin.prompt_studio.batch_runs.create.confirm") } %>
|
|
31
|
+
</div>
|
|
32
|
+
</div>
|
|
33
|
+
</div>
|
|
34
|
+
</div>
|
|
35
|
+
</div>
|