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

# # Que faire quand vous obtenez une erreur
# 
# Ce chapitre portant sur le débogage, la langue nous importe peu ici. Nous nous intéressons surtout à la logique du code pour comprendre d'où provient l'erreur.

# Installez les bibliothèques 🤗 Transformers et 🤗 Datasets pour exécuter ce *notebook*.

# In[ ]:


get_ipython().system('pip install datasets transformers[sentencepiece]')
get_ipython().system('apt install git-lfs')


# Vous aurez besoin de configurer git, adaptez votre email et votre nom dans la cellule suivante.

# In[ ]:


get_ipython().system('git config --global user.email "you@example.com"')
get_ipython().system('git config --global user.name "Your Name"')


# Vous devrez également être connecté au *Hub* d'Hugging Face. Exécutez ce qui suit et entrez vos informations d'identification.

# In[ ]:


from huggingface_hub import notebook_login

notebook_login()


# In[ ]:


from distutils.dir_util import copy_tree
from huggingface_hub import Repository, snapshot_download, create_repo, get_full_repo_name


def copy_repository_template():
    # Cloner le dépôt et extraire le chemin local
    template_repo_id = "lewtun/distilbert-base-uncased-finetuned-squad-d5716d28"
    commit_hash = "be3eaffc28669d7932492681cd5f3e8905e358b4"
    template_repo_dir = snapshot_download(template_repo_id, revision=commit_hash)
    # Créer un dépôt vide sur le Hub
    model_name = template_repo_id.split("/")[1]
    create_repo(model_name, exist_ok=True)
    # Clonez le dépôt vide
    new_repo_id = get_full_repo_name(model_name)
    new_repo_dir = model_name
    repo = Repository(local_dir=new_repo_dir, clone_from=new_repo_id)
    # Copier les fichiers
    copy_tree(template_repo_dir, new_repo_dir)
    # Pousser sur le Hub
    repo.push_to_hub()


# In[ ]:


from transformers import pipeline

model_checkpoint = get_full_repo_name("distillbert-base-uncased-finetuned-squad-d5716d28")
reader = pipeline("question-answering", model=model_checkpoint)


# In[ ]:


model_checkpoint = get_full_repo_name("distilbert-base-uncased-finetuned-squad-d5716d28")
reader = pipeline("question-answering", model=model_checkpoint)


# In[ ]:


from huggingface_hub import list_repo_files

list_repo_files(repo_id=model_checkpoint)


# In[ ]:


from transformers import AutoConfig

pretrained_checkpoint = "distilbert-base-uncased"
config = AutoConfig.from_pretrained(pretrained_checkpoint)


# In[ ]:


config.push_to_hub(model_checkpoint, commit_message="Add config.json")


# In[ ]:


reader = pipeline("question-answering", model=model_checkpoint, revision="main")

context = r"""
Extractive Question Answering is the task of extracting an answer from a text
given a question. An example of a question answering dataset is the SQuAD
dataset, which is entirely based on that task. If you would like to fine-tune a
model on a SQuAD task, you may leverage the
examples/pytorch/question-answering/run_squad.py script.

🤗 Transformers is interoperable with the PyTorch, TensorFlow, and JAX
frameworks, so you can use your favourite tools for a wide variety of tasks!
"""

question = "What is extractive question answering?"
reader(question=question, context=context)


# In[ ]:


tokenizer = reader.tokenizer
model = reader.model


# In[ ]:


question = "Which frameworks can I use?"


# In[ ]:


import torch

inputs = tokenizer(question, context, add_special_tokens=True)
input_ids = inputs["input_ids"][0]
outputs = model(**inputs)
answer_start_scores = outputs.start_logits
answer_end_scores = outputs.end_logits
# Obtenir le début de réponse le plus probable avec l'argmax du score
answer_start = torch.argmax(answer_start_scores)
# Obtenir la fin de réponse la plus probable avec l'argmax du score
answer_end = torch.argmax(answer_end_scores) + 1
answer = tokenizer.convert_tokens_to_string(
    tokenizer.convert_ids_to_tokens(input_ids[answer_start:answer_end])
)
print(f"Question: {question}")
print(f"Answer: {answer}")


# In[ ]:


inputs["input_ids"][:5]


# In[ ]:


type(inputs["input_ids"])