#!/usr/bin/env python
# coding: utf-8

# In[1]:


get_ipython().run_line_magic('pylab', 'inline')
# only display first 30 seconds
from pyannote.core.notebook import set_notebook_crop
from pyannote.core import Segment
set_notebook_crop(Segment(0, 30)) 


# # The Big Bang Theory TVD Plugin

# ## Install

# The following command will install **The Big Bang Theory TVD plugin** (and **TVD** if it is missing)
# ```bash
# pip install TVDTheBigBangTheory
# ```

# ## Download all resources

# The following command will download all resources for **The Big Bang Theory** into `/tmp/` directory.
# ```bash
# python -m tvd.create metadata /tmp/ TheBigBangTheory
# ```

# ## Available resources

# Initialize TheBigBangTheory TVD plugin

# In[2]:


from tvd import TheBigBangTheory
theBigBangTheory = TheBigBangTheory('/tmp')


# Get first episode

# In[3]:


episode = theBigBangTheory.episodes[0]
episode


# ### Outlines

# Source: http://wiki.the-big-bang-theory.com  
# Provides 
#   - scene summary
#   - scene location

# In[9]:


outline = theBigBangTheory.get_resource('outline', episode)
outline


# ### Transcripts

# Source: http://foreverdreaming.org/  
# Provides:
# 
#   - speaker label
#   - speech content
# 
# Does not provide timestamps.

# In[6]:


transcript = theBigBangTheory.get_resource('transcript', episode)
transcript


# ### Force-aligned transcripts

# Source: [LIMSI](http://www.limsi.fr)  
# Provides:
# 
#   - word-level timestamps (start & end time)

# In[7]:


transcript_aligned = theBigBangTheory.get_resource('transcript_aligned', episode)
transcript_aligned.crop(0, 30)


# ### Speaker

# Source: [Makarand Tapaswi](https://cvhci.anthropomatik.kit.edu/~mtapaswi/projects/personid.html)  
# Provides:
# 
#   - speaker label with manual timestamps (five main characters + *other*)

# In[10]:


speaker = theBigBangTheory.get_resource('speaker', episode)
speaker