Module: OllamaChat::PromptManagement
- Included in:
- Chat
- Defined in:
- lib/ollama_chat/prompt_management.rb
Overview
Provides administrative and interactive management for prompt templates stored in the database.
This module handles the user-facing selection process for prompts, allowing users to interactively pick a prompt from the database.
Instance Method Summary collapse
-
#add_new_prompt(context: nil) ⇒ Boolean?
Interactively prompts the user for a name and content (optionally loading from a file) to create a new prompt template.
-
#all_prompts(default: nil, context: nil) ⇒ Array<SearchUI::Wrapper>
Retrieves all stored prompts, decorated with a heart if they are marked as favourites.
-
#choose_and_delete_prompt(context: nil, force: false) ⇒ Object
Interactively selects an existing non-default prompt and deletes it after confirmation.
-
#choose_and_edit_prompt(context: nil) ⇒ self?
Interactively selects an existing prompt and allows the user to edit its content via the integrated editor.
-
#choose_prompt(default: nil, context: nil, prompt: "Select a #{context || 'prompt'} template: %s") ⇒ OllamaChat::Database::Models::Prompt?
The choose_prompt method presents a menu of available prompts for selection.
-
#duplicate_prompt(context: nil) ⇒ self?
Duplicates an existing prompt.
-
#export_prompt(context: nil) ⇒ self?
Interactively exports a prompt to a specified file.
-
#import_prompt(filename, context: nil) ⇒ self?
Interactively imports a prompt from a file.
-
#info_prompt(context: nil) ⇒ self?
Displays detailed information about a selected prompt template.
-
#list_prompts(context: nil) ⇒ Array
Lists all prompt templates in the database, indicating which are defaults and showing a truncated preview of their content.
-
#prepare_conversation_history ⇒ String
Aggregates the current conversation history into a single string for context-aware generation.
-
#reset_prompt_to_default(name, context: nil) ⇒ Boolean?
Resets a prompt's content to the default value defined in the configuration.
-
#suggest_prompts(edit: false) ⇒ String?
Interactively generates follow-up prompt suggestions based on the current session.
Instance Method Details
#add_new_prompt(context: nil) ⇒ Boolean?
Interactively prompts the user for a name and content (optionally loading from a file) to create a new prompt template.
72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 |
# File 'lib/ollama_chat/prompt_management.rb', line 72 def add_new_prompt(context: nil) context ||= 'prompt' switch_history(:add_prompt) do name = determine_valid_new_name_for_prompt('to add', context:) or return sources = %w[ [CLIPBOARD] [FILES] [EMPTY/MANUAL] ] chosen_source = choose_entry(sources, prompt: 'Where shall we source the prompt from? %s') chosen_source or return content = case chosen_source when '[CLIPBOARD]' perform_paste_from_clipboard(edit: false) when '[FILES]' patterns = ask?( prompt: "❓ Enter file patterns to load file, C-u ⇒ new, C-c ⇒ cancel: ", prefill: '**/*.{txt,md}' ) patterns.nil? ? (return) : (patterns.present? ? load_prompt_from_file(patterns) : nil) else nil end prompt_content = edit_text(content) store_prompt(name, prompt_content, context:) log(:info, "Prompt added", data: { name:, context: }) true end end |
#all_prompts(default: nil, context: nil) ⇒ Array<SearchUI::Wrapper>
Retrieves all stored prompts, decorated with a heart if they are marked as favourites.
16 17 18 19 20 21 22 |
# File 'lib/ollama_chat/prompt_management.rb', line 16 def all_prompts(default: nil, context: nil) context ||= 'prompt' favs = all_favourited(context) each_prompt(context:, default:).sort_by(&:name).map do |p| prompt_with_favourite(p.name, favs[p.name]) end end |
#choose_and_delete_prompt(context: nil, force: false) ⇒ Object
Interactively selects an existing non-default prompt and deletes it after confirmation.
103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 |
# File 'lib/ollama_chat/prompt_management.rb', line 103 def choose_and_delete_prompt(context: nil, force: false) context ||= 'prompt' selected_prompt = choose_prompt( default: force ? nil : false, context:, prompt: 'Which template has outlived its usefulness? %s' ) or return STDOUT.puts kramdown_ansi_parse( selected_prompt.to_s + "\n---" ) confirm?( prompt: "🔔 Really delete the prompt #{bold{selected_prompt.name}}? (y/n) ", yes: /\Ay/i ) or return selected_prompt.destroy log(:info, "Prompt deleted", data: { name: selected_prompt.name, context: }) end |
#choose_and_edit_prompt(context: nil) ⇒ self?
Interactively selects an existing prompt and allows the user to edit its content via the integrated editor.
125 126 127 128 129 130 131 132 |
# File 'lib/ollama_chat/prompt_management.rb', line 125 def choose_and_edit_prompt(context: nil) context ||= 'prompt' selected_prompt = choose_prompt(context:, prompt: 'Which spell needs some fine-tuning? %s') or return selected_prompt.['content'] = edit_text(selected_prompt.['content'].to_s) selected_prompt.save log(:info, "Prompt edited", data: { name: selected_prompt.name, context: }) self end |
#choose_prompt(default: nil, context: nil, prompt: "Select a #{context || 'prompt'} template: %s") ⇒ OllamaChat::Database::Models::Prompt?
The choose_prompt method presents a menu of available prompts for selection. It retrieves the list of prompt names from the database, adds an '[EXIT]' option, and displays them via the Chooser utility.
34 35 36 37 38 39 40 41 42 43 44 45 |
# File 'lib/ollama_chat/prompt_management.rb', line 34 def choose_prompt(default: nil, context: nil, prompt: "Select a #{context || 'prompt'} template: %s") context ||= 'prompt' prompts = all_prompts(default:, context:) prompts.unshift('[EXIT]') case chosen = choose_entry(prompts, prompt:) when '[EXIT]', nil STDOUT.puts "Exiting chooser." return when SearchUI::Wrapper prompt(chosen.value, context:) end end |
#duplicate_prompt(context: nil) ⇒ self?
Duplicates an existing prompt.
This method initiates an interactive workflow:
- Prompts the user to select a prompt to clone.
- Displays the content of the selected prompt for verification.
- Requests a new name for the duplicate, validating that it does not already exist in the database.
- Creates and saves the new prompt record using the Database::Duplicatable mixin.
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 |
# File 'lib/ollama_chat/prompt_management.rb', line 145 def duplicate_prompt(context: nil) context ||= 'prompt' selected_prompt = choose_prompt(context:, prompt: 'Which prompt shall be the basis for a new one? %s') or return STDOUT.puts kramdown_ansi_parse( selected_prompt.to_s + "\n---" ) name = nil loop do name = ask?( prompt: "❓ Enter new prompt name to duplicate as, C-c ⇒ cancel: " ) if name.nil? STDOUT.puts "Cancelled." return nil end if prompt(name, context:) STDOUT.puts "Prompt named #{bold{name}} already exists." else break end end duplicated_prompt = selected_prompt.duplicate duplicated_prompt.name = name duplicated_prompt.['default'] = false duplicated_prompt.save log(:info, "Prompt duplicated", data: { name:, old_name: selected_prompt.name, context: }) self end |
#export_prompt(context: nil) ⇒ self?
Interactively exports a prompt to a specified file.
The process follows these steps:
- Prompts the user to select a prompt via
choose_prompt. - Displays the prompt's current content to the terminal.
- Prompts for a destination filename via
determine_valid_output_filename. - Writes the prompt content to the chosen file.
220 221 222 223 224 225 226 227 228 229 230 231 |
# File 'lib/ollama_chat/prompt_management.rb', line 220 def export_prompt(context: nil) context ||= 'prompt' selected_prompt = choose_prompt(context:, prompt: 'Which template are you exporting to disk? %s') or return STDOUT.puts kramdown_ansi_parse( selected_prompt.to_s + "\n---" ) filename = determine_valid_output_filename('to write to') or return filename.write(selected_prompt.to_s) log(:info, "Prompt exported", data: { name: selected_prompt.name, dest: filename.to_s, context: }) STDOUT.puts "Prompt #{selected_prompt.name.inspect} was exported as #{filename.to_path.inspect}?" self end |
#import_prompt(filename, context: nil) ⇒ self?
Interactively imports a prompt from a file.
The process follows these steps:
- Resolves the source file path (either using the provided filename or prompting the user to choose one).
- Prompts for a unique name for the new prompt via
determine_valid_new_name_for_prompt. - Reads the file content and stores it in the database.
186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 |
# File 'lib/ollama_chat/prompt_management.rb', line 186 def import_prompt(filename, context: nil) context ||= 'prompt' if filename if File.exist?(filename) filename = Pathname.new(filename) else filename = choose_filename(filename) end else filename = choose_filename('**/*.md') end unless filename STDOUT.puts "Cancelled." return end prompt_name = determine_valid_new_name_for_prompt('to import', context:) or return prompt_content = filename.read store_prompt(prompt_name, prompt_content, context:) log(:info, "Prompt imported", data: { name: prompt_name, source: filename.to_s, context: }) STDOUT.puts "Imported prompt as #{prompt_name.inspect}." self end |
#info_prompt(context: nil) ⇒ self?
Displays detailed information about a selected prompt template.
50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 |
# File 'lib/ollama_chat/prompt_management.rb', line 50 def info_prompt(context: nil) context ||= 'prompt' if selected_prompt = choose_prompt(context:, prompt: 'Which blueprint would you like to inspect? %s') use_pager do |output| output.puts kramdown_ansi_parse(<<~EOT) # Prompt #{selected_prompt.name} --- #{selected_prompt.to_s} --- EOT end end self end |
#list_prompts(context: nil) ⇒ Array
Lists all prompt templates in the database, indicating which are defaults and showing a truncated preview of their content.
295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 |
# File 'lib/ollama_chat/prompt_management.rb', line 295 def list_prompts(context: nil) context ||= 'prompt' favs = all_favourited(context) each_prompt(context:).sort_by(&:name).map do |p| default = p.['default'] ? '⛭' : '✎' start = '%s %s' % [ default, bold { p.name } ] start = prefix_favourite(start, favs[p.name]) content = p.to_s.inspect[1..-2] content = Kramdown::ANSI::Width.truncate( content, length: 0.9 * (Tins::Terminal.columns - start.size) ) STDOUT.print start STDOUT.puts ' %s' % italic { content } end end |
#prepare_conversation_history ⇒ String
Aggregates the current conversation history into a single string for context-aware generation.
Each message is formatted as "Sender Name: Message Content", skipping messages that contain no content.
240 241 242 243 244 245 246 |
# File 'lib/ollama_chat/prompt_management.rb', line 240 def prepare_conversation_history ..inject('') do |result, | = .content.full? or next result sender_name = sender_name_displayed(, template: false) result << "%s: %s" % [ sender_name, ] end end |
#reset_prompt_to_default(name, context: nil) ⇒ Boolean?
Resets a prompt's content to the default value defined in the configuration.
315 316 317 318 319 320 321 |
# File 'lib/ollama_chat/prompt_management.rb', line 315 def reset_prompt_to_default(name, context: nil) context ||= 'prompt' if content = config.prompts.prompt[name.to_s] store_prompt(name, content, context:) true end end |
#suggest_prompts(edit: false) ⇒ String?
Interactively generates follow-up prompt suggestions based on the current session.
This method constructs a prompt containing the conversation history and an instruction (either selected from a template or provided manually) and requests a generation from the AI model. The resulting suggestions are then opened in the editor for final refinement before being returned.
261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 |
# File 'lib/ollama_chat/prompt_management.rb', line 261 def suggest_prompts(edit: false) instruction = nil if edit # Let the user write a suggestion instruction on the fly instruction = edit_text('').full? or return else # Let the user pick a prompt template (e.g., suggest_coding, suggest_roleplaying) instruction = choose_prompt( prompt: 'Which suggestion strategy shall we employ? %s', context: 'suggest' ) or return end # Build the context by gathering all current conversation messages history = prepare_conversation_history full_prompt = <<~EOT Conversation History: #{history} Instruction: #{instruction} EOT # Execute a silent chat oneshot call (doesn't add to history) suggestions = generate(prompt: full_prompt).full? or return # Pass the AI's suggestions through the editor for final refinement edit_text(suggestions) end |