Merge pull request #102 from fractalego/interrupt-speech

Interrupt speech

documentation/source/index.rst

@@ -16,6 +16,7 @@ Welcome to WAFL's 0.0.90 documentation!
    configuration
    running_WAFL
    facts_and_rules
+   indexing_files
    modify_the_prompt
    examples
    testcases

documentation/source/indexing_files.rst

@@ -0,0 +1,16 @@
+Indexing Files
+==============
+
+There is a special file called `indices.yaml` that is used to define the paths for the files
+that need to be indexed.
+The indexed files can be queried from the interface in a RAG fashion.
+At the moment, the only supported file types are `txt` and `pdf`.
+
+.. code-block:: yaml
+  paths:
+    - /path/to/files/
+
+From the command line you can manually add a path as in the following
+
+.. code-block:: bash
+  $ wafl add /path/to/files/

requirements.txt

@@ -1,21 +1,17 @@
-flask[async]==2.0.1
-flask-cors==3.0.10
-flask_dropzone==1.6.0
-nltk==3.6.2
-gensim==4.3.1
+flask[async]==3.0.3
+flask-cors==4.0.1
+nltk==3.8.1
+gensim==4.3.3
 sklearn==0.0
-python-Levenshtein==0.12.2
-wave==0.0.2
+python-Levenshtein==0.25.1
 fuzzywuzzy==0.18.0
-PyAudio==0.2.13
-num2words==0.5.12
+PyAudio==0.2.14
+num2words==0.5.13
 word2number==1.1
-aiohttp==3.8.4
-werkzeug==2.1.2
-sphinx==6.1.3
-sphinx-rtd-theme==1.2.0
-bluepy==1.3.0
-einops==0.6.1
+aiohttp==3.9.5
+werkzeug==3.0.3
+sphinx==7.4.6
+sphinx-rtd-theme==2.0.0
 g2p-en==2.1.0
 pyyaml==6.0.1
 joblib==1.4.2

setup.py

@@ -20,6 +20,7 @@
         "wafl.connectors.clients",
         "wafl.connectors.factories",
         "wafl.connectors.remote",
+        "wafl.dataclasses",
         "wafl.events",
         "wafl.extractors",
         "wafl.handlers",
@@ -29,6 +30,7 @@
         "wafl.listener",
         "wafl.logger",
         "wafl.parsing",
+        "wafl.readers",
         "wafl.retriever",
         "wafl.runners",
         "wafl.scheduler",
@@ -44,22 +46,20 @@
         ],
     },
     install_requires=[
-        "flask[async]==2.0.1",
-        "flask-cors==3.0.10",
-        "flask_dropzone==1.6.0",
-        "werkzeug==2.1.2",
-        "nltk==3.6.2",
-        "gensim==4.3.1",
+        "flask[async]==3.0.3",
+        "flask-cors==4.0.1",
+        "nltk==3.8.1",
+        "gensim==4.3.3",
         "sklearn==0.0",
-        "python-Levenshtein==0.12.2",
-        "wave==0.0.2",
+        "python-Levenshtein==0.25.1",
         "fuzzywuzzy==0.18.0",
-        "PyAudio==0.2.13",
-        "pyctcdecode==0.2.1",
-        "num2words==0.5.12",
+        "PyAudio==0.2.14",
+        "num2words==0.5.13",
         "word2number==1.1",
-        "aiohttp==3.8.4",
-        "einops==0.6.1",
+        "aiohttp==3.9.5",
+        "werkzeug==3.0.3",
+        "sphinx==7.4.6",
+        "sphinx-rtd-theme==2.0.0",
         "g2p-en==2.1.0",
         "pyyaml==6.0.1",
         "joblib==1.4.2",

todo.txt

@@ -1,7 +1,11 @@
-* interruptible speech
 * dependabot!!!
+* update readme with index.
+
 * use poetry
 
+
+/* interruptible speech
+
 * upload to hetzner and make it work for some retrieval tasks
 * develop more rules + use-cases for voice and other
 

wafl/command_line.py

@@ -28,9 +28,11 @@ def print_help():
     print("> wafl run-audio: Run a voice-powered version of the chatbot")
     print("> wafl run-server: Run a webserver version of the chatbot")
     print("> wafl run-tests: Run the tests in testcases.txt")
+    print("> wafl add <PATH>: Add the file or folder at <PATH> to the index")
     print(
         "> wafl run-action <ACTION_NAME>: Run the action <ACTION_NAME> from actions.yaml"
     )
+    print("> wafl help: Show this help message")
     print()
 
 
@@ -50,7 +52,7 @@ def process_cli():
             create_initial_files()
             download_models()
 
-        if command == "run":
+        elif command == "run":
             run_app()
             remove_preprocessed("/")
 

wafl/interface/list_interface.py

@@ -17,10 +17,10 @@ async def output(self, text: str, silent: bool = False):
         )
 
     async def input(self) -> str:
-        done, pending = await asyncio.wait(
-            [interface.input() for interface in self._interfaces_list],
-            return_when=asyncio.FIRST_COMPLETED,
-        )
+        tasks = [interface.input() for interface in self._interfaces_list]
+        done, pending = await asyncio.wait(tasks, return_when=asyncio.FIRST_COMPLETED)
+        for task in pending:
+            task.cancel()
         return done.pop().result()
 
     async def insert_input(self, text: str):

wafl/interface/voice_interface.py

@@ -1,9 +1,8 @@
+import asyncio
 import os
 import random
 import re
 
-import nltk
-
 from wafl.events.utils import remove_text_between_brackets
 from wafl.interface.base_interface import BaseInterface
 from wafl.interface.utils import not_good_enough
@@ -57,14 +56,15 @@ async def output(self, text: str, silent: bool = False):
         text = text
         self._insert_utterance(speaker="bot", text=text)
         print(COLOR_START + "bot> " + text + COLOR_END)
-        for sentence in nltk.sent_tokenize(text):
-            await self._speaker.speak(sentence)
+        await self._speaker.speak(text)
         self.bot_has_spoken(True)
 
     async def input(self) -> str:
         text = ""
         while not text:
             text = await self._listener.input()
+            if not text.strip():
+                continue
             text = self.__remove_activation_word_and_normalize(text)
             hotword = await self._listener.get_hotword_if_present()
             if hotword:

wafl/knowledge/indexing_implementation.py

@@ -10,6 +10,9 @@
 
 async def _add_indices_to_knowledge(knowledge, text):
     indices = yaml.safe_load(text)
+    if "paths" not in indices or not indices["paths"]:
+        return knowledge
+
     for path in indices["paths"]:
         for root, _, files in os.walk(path):
             for file in files:

wafl/runners/routes.py

@@ -1,6 +1,6 @@
 import os
 
-from flask import Flask
+from flask import Flask, abort
 from flask_cors import CORS
 
 _path = os.path.dirname(__file__)
@@ -12,75 +12,101 @@
 )
 CORS(app)
 
+_routes_dict = {}
+
 
 def get_app():
     return app
 
 
-def add_new_rules(app: Flask, conversation_id: int, web_server_loop: "WebLoop"):
-    app.add_url_rule(
-        f"/{conversation_id}/",
-        f"index_{conversation_id}",
-        web_server_loop.index,
-        methods=["GET"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/reset_conversation",
-        f"reset_conversation_{conversation_id}",
-        web_server_loop.reset_conversation,
-        methods=["POST"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/reload_rules",
-        f"reload_rules_{conversation_id}",
-        web_server_loop.reload_rules,
-        methods=["POST"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/check_new_messages",
-        f"check_new_messages_{conversation_id}",
-        web_server_loop.check_for_new_messages,
-        methods=["POST"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/load_messages",
-        f"load_messages_{conversation_id}",
-        web_server_loop.load_messages,
-        methods=["POST", "GET"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/input",
-        f"input_{conversation_id}",
-        web_server_loop.handle_input,
-        methods=["POST", "GET"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/output",
-        f"output_{conversation_id}",
-        web_server_loop.handle_output,
-        methods=["POST"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/thumbs_up",
-        f"thumbs_up_{conversation_id}",
-        web_server_loop.thumbs_up,
-        methods=["POST"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/thumbs_down",
-        f"thumbs_down_{conversation_id}",
-        web_server_loop.thumbs_down,
-        methods=["POST"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/toggle_logs",
-        f"toggle_logs_{conversation_id}",
-        web_server_loop.toggle_logs,
-        methods=["POST"],
-    )
-    app.add_url_rule(
-        f"/{conversation_id}/get_info",
-        f"get_info_{conversation_id}",
-        web_server_loop.get_info,
-        methods=["POST"],
-    )
+@app.route("/<conversation_id>/index", methods=["GET"])
+async def get_index(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["index"]()
+
+
+@app.route("/<conversation_id>/reset_conversation", methods=["POST"])
+async def reset_conversation(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["reset_conversation"]()
+
+
+@app.route("/<conversation_id>/reload_rules", methods=["POST"])
+async def reload_rules(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["reload_rules"]()
+
+
+@app.route("/<conversation_id>/check_new_messages", methods=["POST"])
+async def check_new_messages(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["check_new_messages"]()
+
+
+@app.route("/<conversation_id>/load_messages", methods=["POST", "GET"])
+async def load_messages(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["load_messages"]()
+
+
+@app.route("/<conversation_id>/input", methods=["POST", "GET"])
+async def input(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["input"]()
+
+
+@app.route("/<conversation_id>/output", methods=["POST"])
+async def output(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["output"]()
+
+
+@app.route("/<conversation_id>/thumbs_up", methods=["POST"])
+async def thumbs_up(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["thumbs_up"]()
+
+
+@app.route("/<conversation_id>/thumbs_down", methods=["POST"])
+async def thumbs_down(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["thumbs_down"]()
+
+
+@app.route("/<conversation_id>/toggle_logs", methods=["POST"])
+async def toggle_logs(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["toggle_logs"]()
+
+
+@app.route("/<conversation_id>/get_info", methods=["POST"])
+async def get_info(conversation_id: str):
+    if conversation_id not in _routes_dict:
+        abort(404)
+    return await _routes_dict[conversation_id]["get_info"]()
+
+
+def add_new_routes(conversation_id: str, web_server_handler: "WebHandler"):
+    _routes_dict[str(conversation_id)] = {
+        "index": web_server_handler.index,
+        "reset_conversation": web_server_handler.reset_conversation,
+        "reload_rules": web_server_handler.reload_rules,
+        "check_new_messages": web_server_handler.check_for_new_messages,
+        "load_messages": web_server_handler.load_messages,
+        "input": web_server_handler.handle_input,
+        "output": web_server_handler.handle_output,
+        "thumbs_up": web_server_handler.thumbs_up,
+        "thumbs_down": web_server_handler.thumbs_down,
+        "toggle_logs": web_server_handler.toggle_logs,
+        "get_info": web_server_handler.get_info,
+    }