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

# This notebook contains an example of combining checkpoint use inside a input function in Snakemake 8.18.2. The notebook was based on a [work](https://edwards.flinders.edu.au/how-to-use-snakemake-checkpoints/)  written on an outdated version of the library.
# 
# Acknowledgments: I would like to thank [Wayne](https://stackoverflow.com/users/8508004/wayne) for his help in implementation.

# In[1]:


get_ipython().system(' snakemake --version')


# # Case

# In[1]:


get_ipython().system(' cat Snakefile')


# In the source code, the {SAMPLE} wildcard uses the value of the SMP variable. However, SMP is initially defined as None, which may cause the code to not work correctly. When a checkpoint is executed, the SMP variable changes and must maintain a list of file names in the OUTDIR directories.

# In[3]:


get_ipython().system('snakemake -c 1 --debug-dag')


# The logs show that after executing make_five_files, the {SAMPLE} wildcard gets the correct values (for example, 10536, 16393, etc.). copy_files and list_all_files are selected multiple times with SAMPLE=None, which is not correct behavior.

# In[5]:


get_ipython().system('ls')


# In[2]:


get_ipython().system('ls first_directory/')


# In[4]:


get_ipython().system('ls second_directory/')


# # Solution

# Using a function that returns a list of file names from a directory. The OUTDIR directory is used as input in the all rule.

# In[6]:


get_ipython().system(' cat Snakefile')


# In[2]:


get_ipython().system(' snakemake -c 1 -F')


# In[3]:


get_ipython().system(' ls')


# In[4]:


get_ipython().system(' ls first_directory/')


# Let's add the creation of files in a new directory. The directory name is specified by the SNDDIR variable at the beginning of the file.

# In[7]:


get_ipython().system(' cat Snakefile')


# In[14]:


get_ipython().system(' snakemake -c 1 -F')


# In[15]:


get_ipython().system(' ls')


# In[16]:


get_ipython().system(' ls  first_directory/')


# In[17]:


get_ipython().system(' ls second_directory/')


# In[ ]: