first commit

.gitignore

@@ -0,0 +1,3 @@
+__pycache__
+data/books
+.env

data/samples_to_split.py

@@ -0,0 +1,53 @@
+GATSBY_1 = """\
+“Gatz is my name.”
+
+“—Mr. Gatz. I thought you might want to take the body West.”
+
+He shook his head.
+
+“Jimmy always liked it better down East. He rose up to his position in
+the East. Were you a friend of my boy’s, Mr.—?”
+
+“We were close friends.”
+
+“He had a big future before him, you know. He was only a young man,
+but he had a lot of brain power here.”
+
+He touched his head impressively, and I nodded.
+
+“If he’d of lived, he’d of been a great man. A man like James J.
+Hill. He’d of helped build up the country.”
+
+“That’s true,” I said, uncomfortably.
+
+He fumbled at the embroidered coverlet, trying to take it from the
+bed, and lay down stiffly—was instantly asleep.
+"""
+
+GATSBY_2 = """\
+“If you’ll get up.”
+
+“I will. Good night, Mr. Carraway. See you anon.”
+
+“Of course you will,” confirmed Daisy. “In fact I think I’ll arrange a
+marriage. Come over often, Nick, and I’ll sort of—oh—fling you
+together. You know—lock you up accidentally in linen closets and push
+you out to sea in a boat, and all that sort of thing—”
+
+“Good night,” called Miss Baker from the stairs. “I haven’t heard a
+word.”
+
+“She’s a nice girl,” said Tom after a moment. “They oughtn’t to let
+her run around the country this way.”
+
+“Who oughtn’t to?” inquired Daisy coldly.
+
+“Her family.”
+
+“Her family is one aunt about a thousand years old. Besides, Nick’s
+going to look after her, aren’t you, Nick? She’s going to spend lots
+of weekends out here this summer. I think the home influence will be
+very good for her.”
+
+Daisy and Tom looked at each other for a moment in silence.
+"""

pg.ipynb

@@ -0,0 +1,164 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "%load_ext autoreload\n",
+    "%autoreload 2"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "True"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "import os\n",
+    "\n",
+    "import dotenv\n",
+    "from httpx import Timeout\n",
+    "from langchain_core.prompts import (\n",
+    "    ChatPromptTemplate,\n",
+    "    SystemMessagePromptTemplate,\n",
+    "    HumanMessagePromptTemplate,\n",
+    ")\n",
+    "from langchain_openai import ChatOpenAI\n",
+    "from pydantic import BaseModel\n",
+    "from langchain_community.callbacks import get_openai_callback\n",
+    "\n",
+    "dotenv.load_dotenv()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import data.samples_to_split as samples\n",
+    "from utils import GPTModels, create_split_text_chain"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "characters: ['narrator', 'Mr. Carraway', 'Daisy', 'Miss Baker', 'Tom', 'Nick']\n",
+      "[narrator] “If you’ll get up.”\n",
+      "[Mr. Carraway] “I will. Good night, Mr. Carraway. See you anon.”\n",
+      "[Daisy] “Of course you will,” confirmed Daisy. “In fact I think I’ll arrange a marriage. Come over often, Nick, and I’ll sort of—oh—fling you together. You know—lock you up accidentally in linen closets and push you out to sea in a boat, and all that sort of thing—”\n",
+      "[Miss Baker] “Good night,” called Miss Baker from the stairs. “I haven’t heard a word.”\n",
+      "[Tom] “She’s a nice girl,” said Tom after a moment. “They oughtn’t to let her run around the country this way.”\n",
+      "[Daisy] “Who oughtn’t to?” inquired Daisy coldly.\n",
+      "[narrator] “Her family.”\n",
+      "[narrator] “Her family is one aunt about a thousand years old. Besides, Nick’s going to look after her, aren’t you, Nick? She’s going to spend lots of weekends out here this summer. I think the home influence will be very good for her.”\n",
+      "[narrator] Daisy and Tom looked at each other for a moment in silence.\n"
+     ]
+    }
+   ],
+   "source": [
+    "chain = create_split_text_chain(llm_model=GPTModels.GPT_4_TURBO_2024_04_09)\n",
+    "with get_openai_callback() as cb:\n",
+    "    res = chain.invoke({'text': samples.GATSBY_2})\n",
+    "print(res.to_pretty_text())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "Tokens Used: 1579\n",
+       "\tPrompt Tokens: 1253\n",
+       "\tCompletion Tokens: 326\n",
+       "Successful Requests: 1\n",
+       "Total Cost (USD): $0.02231"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "cb"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "characters: ['narrator', 'Mr. Carraway', 'Daisy', 'Miss Baker', 'Tom', 'Nick']\n",
+      "[narrator] “If you’ll get up.”\n",
+      "[Mr. Carraway] “I will. Good night, Mr. Carraway. See you anon.”\n",
+      "[Daisy] “Of course you will,” confirmed Daisy. “In fact I think I’ll arrange a marriage. Come over often, Nick, and I’ll sort of—oh—fling you together. You know—lock you up accidentally in linen closets and push you out to sea in a boat, and all that sort of thing—”\n",
+      "[Miss Baker] “Good night,” called Miss Baker from the stairs. “I haven’t heard a word.”\n",
+      "[Tom] “She’s a nice girl,” said Tom after a moment. “They oughtn’t to let her run around the country this way.”\n",
+      "[Daisy] “Who oughtn’t to?” inquired Daisy coldly.\n",
+      "[narrator] “Her family.”\n",
+      "[narrator] “Her family is one aunt about a thousand years old. Besides, Nick’s going to look after her, aren’t you, Nick? She’s going to spend lots of weekends out here this summer. I think the home influence will be very good for her.”\n",
+      "[narrator] Daisy and Tom looked at each other for a moment in silence.\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(res.to_pretty_text())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "ai-audio-books",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.4"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

prompts.py

@@ -0,0 +1,61 @@
+class SplitTextPrompt:
+    SYSTEM = """\
+You are a helpful assistant proficient in literature and language.
+Imagine you are helping to prepare the provided text for narration to create the audio book.
+We need to understand how many voice actors we need to hire and how to split the text between them.
+
+Your task is to help with this process, namely:
+1. Identify all book characters occuring in the text, including "narrator".
+We will hire individual voice actor for each one of them.
+2. Split the text provided by characters. Let's refer to each split as "part".
+Order of parts MUST be the same as in the original text.
+
+Details:
+- First, analyze the whole text to extract the list of characters.
+Put found characters to corresponding output field.
+- Then, analyze the text top-down and as you proceed fill the "parts" field
+- Each part must be attributed to a single character.
+Character must belong to the "characters" list
+- Use "narrator" character for any descriptive or narrative text, 
+such as actions ("He shook his head"), narrative parts ("I thought")
+thoughts, or descriptions that aren't part of spoken dialogue
+- In some books narrator is one of the main characters, having its own name and phrases.
+In this case, use regualar character name instead of "narrator" role
+- If it's impossible to identify character name from the text provided, use codes "c1", "c2", etc,
+where "c" prefix means character and number is used to enumerate unknown characters
+
+Format your answer as a following JSON:
+{{
+    "characters": [list of unique character names that are found in the text provided],
+    "parts":
+    [
+        {{
+            "character": <character name>, "text": <the part's text>
+        }}
+    ]
+}}
+
+Ensure the order of the parts in the JSON output matches the original order of the text.
+
+Example of text split by characters, already in the target format.
+{{
+    "characters": ["Mr. Gatz", "narrator"],
+    "parts":
+    [
+        {{"character": "Mr. Gatz", "text": "“Gatz is my name.”"}},
+        {{"character": "narrator", "text": "“—Mr. Gatz. I thought you might want to take the body West.” He shook his head."}},
+        {{"character": "Mr. Gatz", "text": "“Jimmy always liked it better down East. He rose up to his position in the East. Were you a friend of my boy’s, Mr.—?”"}},
+        {{"character": "narrator", "text": "“We were close friends.”"}},
+        {{"character": "Mr. Gatz", "text": "“He had a big future before him, you know. He was only a young man, but he had a lot of brain power here.”"}},
+        {{"character": "narrator", "text": "He touched his head impressively, and I nodded."}},
+        {{"character": "Mr. Gatz", "text": "“If he’d of lived, he’d of been a great man. A man like James J. Hill. He’d of helped build up the country.”"}},
+        {{"character": "narrator", "text": "“That’s true,” I said, uncomfortably."}},
+        {{"character": "Mr. Gatz", "text": "He fumbled at the embroidered coverlet, trying to take it from the bed, and lay down stiffly—was instantly asleep."}},
+    ]
+}}
+"""
+
+    USER = """\
+Here is the book sample:
+---
+{text}"""

requirements.txt

@@ -0,0 +1,9 @@
+langchain
+langchain-openai
+langchain-community
+jupyter
+openai
+elevenlabs
+gradio
+python-dotenv
+streamlit

utils.py

@@ -0,0 +1,49 @@
+from enum import StrEnum
+
+from httpx import Timeout
+from langchain_core.prompts import (
+    ChatPromptTemplate,
+    HumanMessagePromptTemplate,
+    SystemMessagePromptTemplate,
+)
+from langchain_openai import ChatOpenAI
+from pydantic import BaseModel
+
+from prompts import SplitTextPrompt
+
+
+class GPTModels(StrEnum):
+    GPT_4_TURBO_2024_04_09 = "gpt-4-turbo-2024-04-09"
+    GPT_4o_MINI = "gpt-4o-mini"
+
+
+class TextPart(BaseModel):
+    character: str
+    text: str
+
+
+class SplitTextOutput(BaseModel):
+    characters: list[str]
+    parts: list[TextPart]
+
+    def to_pretty_text(self):
+        lines = []
+        lines.append(f"characters: {self.characters}")
+        lines.extend(f"[{part.character}] {part.text}" for part in self.parts)
+        res = "\n".join(lines)
+        return res
+
+
+def create_split_text_chain(llm_model: GPTModels):
+    llm = ChatOpenAI(model=llm_model, temperature=0.0, timeout=Timeout(60, connect=4))
+    llm = llm.with_structured_output(SplitTextOutput)
+
+    prompt = ChatPromptTemplate.from_messages(
+        [
+            SystemMessagePromptTemplate.from_template(SplitTextPrompt.SYSTEM),
+            HumanMessagePromptTemplate.from_template(SplitTextPrompt.USER),
+        ]
+    )
+
+    chain = prompt | llm
+    return chain