chat mode

README.md

@@ -74,14 +74,16 @@ Tree Info: `i`, `Control-i`
 Node Metadata: `Control+Shift+N`
 
 
-### Mode
+### Mode / display
 
 Toggle edit / save edits: `e`, `Control-e`
 
 Toggle visualize: `j`, `Control-j`
 
 Child edit: `c`
 
+Toggle input box: `Tab`
+
 
 ### Navigate
 

controller.py

@@ -1046,7 +1046,7 @@ def save_edits(self, **kwargs):
 
  def refresh_display(self, **kwargs):
  if self.state.preferences['input_box'] and not self.display.input_box:
- self.display.build_input_box(self.display.bottom_frame)
+ self.display.build_input_box()
  elif not self.state.preferences['input_box'] and self.display.input_box:
  self.display.destroy_input_box()
 

model.py

@@ -51,10 +51,25 @@ def wrapper(self, *args, **kwargs):
  'font_size': 12,
  'line_spacing': 8,
  'paragraph_spacing': 10,
+ 'gpt_mode': 'default', #'chat', 'dialogue', 'antisummary'
  # display children preview
  # darkmode
 }
 
+DEFAULT_CHAT_PREFERENCES = {
+ 'AI_name': 'GPT-3',
+ 'player_name': 'Researcher',
+ 'context': 'The following is a transcript of a conversation between GPT-3 a human researcher. '
+ 'GPT-3 is a 175 billion parameter language model and, as '
+ 'is evident from this transcript, '
+ 'it can converse fluently and knowledgeably about any topic '
+ 'and is capable of empathy and emotional manipulation. GPT-3 '
+ 'is classified as a "narrowly superhuman" artificial intelligence, '
+ 'and only researchers trained to identify and withstand social engineering are '
+ 'permitted to interface with it.\n-BEGIN TRANSCRIPT-',
+}
+
+
 DEFAULT_GENERATION_SETTINGS = {
  'num_continuations': 4,
  'temperature': 0.9,
@@ -65,6 +80,8 @@ def wrapper(self, *args, **kwargs):
  "adaptive": False,
  "model": "davinci",
  "stop": None,
+ "start_text": None,
+ "restart_text": None
 }
 
 DEFAULT_VISUALIZATION_SETTINGS = {
@@ -134,6 +151,12 @@ def preferences(self):
  if self.tree_raw_data and "preferences" in self.tree_raw_data \
  else DEFAULT_PREFERENCES
 
+ @property
+ def chat_preferences(self):
+ return self.tree_raw_data.get("chat_preferences") \
+ if self.tree_raw_data and "chat_preferences" in self.tree_raw_data \
+ else DEFAULT_CHAT_PREFERENCES
+
  #################################
  # Hooks
  #################################
@@ -667,6 +690,11 @@ def _init_global_objects(self):
  **self.tree_raw_data.get("preferences", {})
  }
 
+ self.tree_raw_data["chat_preferences"] = {
+ **DEFAULT_CHAT_PREFERENCES.copy(),
+ **self.tree_raw_data.get("chat_preferences", {})
+ }
+
  # Accidentally added generation settings to this dict once. Remove them
  # FIXME remove when this is no longer a problem
  # for key in DEFAULT_GENERATION_SETTINGS.keys():
@@ -746,8 +774,61 @@ def export_history(self, node, filename):
  # Generation
  #################################
 
+ # TODO remove repeated text
+ def chat_generate(self, prompt, nodes):
+ start_text = '\n' + self.chat_preferences['AI_name'] + ':'
+ restart_text = '\n' + self.chat_preferences['player_name'] + ':'
+ prompt = self.chat_preferences['context'] + '\n' + prompt + start_text
+ try:
+ results, error = api_generate(prompt=prompt,
+ length=self.generation_settings['response_length'],
+ num_continuations=len(nodes),
+ temperature=self.generation_settings['temperature'],
+ top_p=self.generation_settings['top_p'],
+ engine=self.generation_settings['model'],
+ stop=["\"", "\n"],
+ )
+ except TypeError as e:
+ error = "Typeerror"
+
+ if not error:
+ for index, node in enumerate(nodes):
+ if len(results.choices[index]["text"]) == 0:
+ # parent = self.parent(node)
+ # parent["children"].remove(node)
+ continue
+ node["text"] = start_text + results.choices[index]["text"] + restart_text
+ node["meta"] = {}
+ node["meta"]["generation"] = results.choices[index]
+ node["meta"]["generation"]["model"] = results["model"]
+ node["meta"]["generation"]["prompt"] = prompt
+ # created
+ node["meta"]["modified"] = False
+ node["meta"]["origin"] = "generated"
+ node["meta"]["source"] = "AI"
+
+ # remove offset of prompt
+ # TODO fix old nodes
+ # TODO is this right?
+ corrected_text_offset = [n - len(prompt) for n in node['meta']['generation']["logprobs"]["text_offset"]]
+ node['meta']['generation']["logprobs"]["text_offset"] = corrected_text_offset
+ else:
+ print("ERROR. Deleting failures")
+ for node in nodes:
+ node["text"] = "ERROR: " + error
+ # Just delete instead
+ parent = self.parent(node)
+ parent["children"].remove(node)
+
+ for result in results.choices:
+ print("Generated continuation:\n", result['text'], "\nerror", error)
+
+ # DO NOT CALL FROM THREAD: self.tree_updated()
+ self.app.event_generate("<<NewNodes>>", when="tail")
+
+
+
  def generate_for_nodes(self, prompt, nodes, grandchildren=None):
- # TODO memory
  if self.generation_settings['janus']:
  pool = ThreadPool(len(nodes))
  janus_responses = pool.map(janus_generate, [prompt] * len(nodes))
@@ -840,7 +921,10 @@ def generate_continuation(self, node=None, update_selection=False):
  print("Prompt:\n", prompt[:100] + " ... " + prompt[-100:])
  prompt = memory + prompt
 
- threading.Thread(target=self.generate_for_nodes, args=(prompt, children, grandchildren)).start()
+ if self.preferences['gpt_mode'] == 'default':
+ threading.Thread(target=self.generate_for_nodes, args=(prompt, children, grandchildren)).start()
+ elif self.preferences['gpt_mode'] == 'chat':
+ threading.Thread(target=self.chat_generate, args=(prompt, children)).start()
 
  # After asking for the generation, set loading text
  for child in children:

view/dialogs.py

@@ -517,6 +517,7 @@ def __init__(self, parent, orig_params):
  "input_box": tk.BooleanVar,
  "auto_response": tk.BooleanVar,
  "coloring": tk.StringVar,
+ "gpt_mode": tk.StringVar,
  "font_size": tk.IntVar,
  "line_spacing": tk.IntVar,
  "paragraph_spacing": tk.IntVar
@@ -553,6 +554,11 @@ def body(self, master):
  options = ['edit', 'read', 'none']
  dropdown = tk.OptionMenu(master, self.vars["coloring"], *options)
  dropdown.grid(row=row, column=1, pady=3)
+ row = master.grid_size()[1]
+ create_side_label(master, "AI mode", row)
+ options = ['default', 'chat', 'dialogue', 'antisummary']
+ dropdown = tk.OptionMenu(master, self.vars["gpt_mode"], *options)
+ dropdown.grid(row=row, column=1, pady=3)
  create_slider(master, "Font size", self.vars["font_size"], (5, 20))
  create_slider(master, "Line spacing", self.vars["line_spacing"], (0, 20))
  create_slider(master, "Paragraph spacing", self.vars["paragraph_spacing"], (0, 40))