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

# # Logfile parsing

# Here, we provide an example concerning how to benefit from QCoDeS logs by simple analysis and visualisation.

# In[1]:


get_ipython().run_line_magic('matplotlib', 'inline')
import os
from zipfile import ZipFile

import matplotlib.pyplot as plt

from qcodes.logger import log_to_dataframe, time_difference


# In[2]:


# put the 30MB into a zip file
filepath = os.path.join(os.getcwd(), "static", "pythonlog.zip")
with ZipFile(filepath) as z:
    with z.open("pythonlog.log", "r") as f:
        my_log = [line.decode() for line in f]


# In[3]:


os.path.exists(filepath)


# In[4]:


logdata = log_to_dataframe(
    my_log,
    separator=" - ",
    columns=["time", "module", "function", "loglevel", "message"],
)


# The `logdata` is, now, a nice and tidy `DataFrame` that one can further manipulate to extract more information, if needed.

# In[5]:


logdata


# In[6]:


data = logdata


# ### Get the query time for the SR860
# 
# We know that the log file documents an experiment quering an SR860 for R and amplitude over and over. Let us analyse and visualise query response times.

# In[7]:


qstr_R = r"Querying instrument SR860_120: OUTP\? 2"  # remember to escape
queries_R = data[data.message.str.contains(qstr_R)]
responses_R = data.loc[queries_R.index + 1]

qstr_lvl = r"Querying instrument SR860_120: SLVL\?"  # remember to escape
queries_lvl = data[data.message.str.contains(qstr_lvl)][:-1]
responses_lvl = data.loc[queries_lvl.index + 1]


# ### Find the elapsed times

# In[8]:


elapsed_times_R = time_difference(queries_R.time, responses_R.time)
elapsed_times_lvl = time_difference(queries_lvl.time, responses_lvl.time)


# ## Visualise!

# In[9]:


from pandas.plotting import register_matplotlib_converters

register_matplotlib_converters()

fig, ax = plt.subplots(1, 1)
ax.plot(
    queries_R.time.str.replace(",", ".").astype("datetime64[ns]"),
    elapsed_times_R,
    ".",
    label="R",
)
ax.plot(
    queries_lvl.time.str.replace(",", ".").astype("datetime64[ns]"),
    elapsed_times_lvl,
    ".",
    label="LVL",
)
fig.autofmt_xdate()
ax.set_ylabel("Response time (s)")
plt.legend()


# In[ ]: