Ralf Magnus Palmiste
f887c94a70
Added the Data, fitter, and statistiline_filter scripts that are necessary working with the model.
193 lines
4.9 KiB
Python
193 lines
4.9 KiB
Python
# -*- coding: utf-8 -*-
|
|
|
|
#%% import
|
|
|
|
import h5py
|
|
import os
|
|
import pandas as pd
|
|
import statsmodels.api as sm
|
|
|
|
from statsmodels.formula.api import ols
|
|
from scipy.stats import f
|
|
from scipy.stats import shapiro
|
|
|
|
h5_file = 'ltcc_current.h5'
|
|
|
|
sobiv_eid_list = []
|
|
ttx_eid_list = []
|
|
teised_eid_list = []
|
|
|
|
with h5py.File(h5_file, 'r') as h5_file:
|
|
|
|
for eid in h5_file.keys():
|
|
if 'tag' in h5_file[eid].attrs:
|
|
tag_val = h5_file[eid].attrs['tag']
|
|
|
|
# pean tõdema, et siin aitas chatgpt oma soovitusega :/ vaga halb debug oli
|
|
if isinstance(tag_val, bytes):
|
|
tag_val = tag_val.decode('utf-8')
|
|
|
|
puhas_eid = eid.replace(" ", "_").replace(":", "-")
|
|
fit_result_eid = "fit_results_" + puhas_eid
|
|
|
|
if tag_val == 'iso':
|
|
sobiv_eid_list.append(fit_result_eid)
|
|
elif tag_val == 'ttx':
|
|
ttx_eid_list.append(fit_result_eid)
|
|
else:
|
|
teised_eid_list.append(fit_result_eid)
|
|
#for attr_key, attr_value in h5_file[eid].attrs.items():
|
|
#print(f"Atribute: {attr_key}, Value: {attr_value}")
|
|
|
|
#%%
|
|
|
|
file = 'ltcc_current.h5'
|
|
|
|
with h5py.File(file, 'r') as h5_file:
|
|
for eid in h5_file.keys():
|
|
puhastatud_eid = eid.replace(" ", "_").replace(":", "-")
|
|
|
|
atribuudid = h5_file[eid].attrs
|
|
sex = atribuudid.get('sex')
|
|
spid = atribuudid.get('spid')
|
|
|
|
csv_file_name = f"fit_results_{puhastatud_eid}.csv"
|
|
|
|
if os.path.exists(csv_file_name):
|
|
df = pd.read_csv(csv_file_name)
|
|
|
|
df['sex'] = sex
|
|
df['spid'] = spid.replace("Mouse AGAT","")
|
|
df['eid'] = eid
|
|
|
|
df.to_csv(csv_file_name, index=False)
|
|
|
|
for fail in os.listdir():
|
|
if fail.endswith('.csv'):
|
|
|
|
eksperiment_id = fail.replace(".csv", "")
|
|
|
|
if eksperiment_id in sobiv_eid_list:
|
|
|
|
df = pd.read_csv(fail)
|
|
df['tag'] = 'iso'
|
|
df.to_csv(fail, index=False)
|
|
|
|
elif eksperiment_id in ttx_eid_list:
|
|
|
|
df = pd.read_csv(fail)
|
|
df['tag'] = 'ttx'
|
|
df.to_csv(fail, index=False)
|
|
|
|
else:
|
|
|
|
df = pd.read_csv(fail)
|
|
df['tag'] = 'teised'
|
|
df.to_csv(fail, index=False)
|
|
|
|
#%%
|
|
|
|
comb_df = pd.DataFrame()
|
|
|
|
for filename in os.listdir():
|
|
|
|
if filename.endswith('.csv'):
|
|
df = pd.read_csv(filename)
|
|
|
|
if 'tag' in df.columns and df['tag'].isin(['iso', 'ttx']).all():
|
|
|
|
comb_df = pd.concat([comb_df, df], ignore_index=True)
|
|
|
|
print(comb_df)
|
|
|
|
#%% lugemine, et teha kindlaks mis tüüpi ANOVA teha kasutades statsmodels packetit,
|
|
|
|
sex_counts = comb_df['sex'].value_counts()
|
|
|
|
spid_counts = comb_df['spid'].value_counts()
|
|
|
|
tag_counts = comb_df['tag'].value_counts()
|
|
|
|
|
|
print(sex_counts,spid_counts,tag_counts)
|
|
|
|
#%% normaalsuse kontroll
|
|
tau_xfer = comb_df['tau_xfer']
|
|
|
|
stat, p = shapiro(tau_xfer)
|
|
|
|
alpha = 0.05
|
|
if p > alpha:
|
|
print("Andmed on normaalselt jaotunud (ei lükka tagasi nullhüpoteesi)")
|
|
else:
|
|
print("Andmed ei ole normaalselt jaotunud (lükata tagasi nullhüpotees)")
|
|
|
|
|
|
|
|
#%% ANOVA 2 WAY
|
|
|
|
comb_df['spid'] = comb_df['spid'].astype('category')
|
|
comb_df['sex'] = comb_df['sex'].astype('category')
|
|
comb_df['tag'] = comb_df['tag'].astype('category')
|
|
|
|
model = ols('tau_xfer ~ C(sex) + C(spid)+ C(tag)', data=comb_df).fit()
|
|
|
|
anova_table = sm.stats.anova_lm(model, typ=2)
|
|
print(anova_table)
|
|
|
|
#%% kriitiline vaartus
|
|
|
|
df_between_groups = 2 # Vabadusastmed gruppide vahel
|
|
df_within_groups = 67 # Vabadusastmed rühmades (Residual)
|
|
|
|
alpha = 0.05
|
|
|
|
critical_f = f.ppf(1 - alpha, df_between_groups, df_within_groups)
|
|
|
|
print("Kriitiline F-väärtus:", critical_f)
|
|
|
|
|
|
#%% kriitiline vaartus
|
|
|
|
|
|
df_between_groups = 2 # Vabadusastmed gruppide vahel (sex ja spid)
|
|
df_within_groups = 36 # Vabadusastmed rühmades (Residual)
|
|
|
|
alpha = 0.05
|
|
|
|
critical_f = f.ppf(1 - alpha, df_between_groups, df_within_groups)
|
|
|
|
print("Kriitiline F-väärtus:", critical_f)
|
|
|
|
"""
|
|
#%%grupeeringud
|
|
|
|
groups = comb_df.groupby('tag')
|
|
|
|
iso_group = groups.get_group('iso')
|
|
ttx_group = groups.get_group('ttx')
|
|
|
|
print("ISO grupp:", iso_group)
|
|
print("nTTX grupp:", ttx_group)
|
|
|
|
#%% ANOVA jaoks on grupid piisavalt suured,
|
|
#allikas https://support.minitab.com/en-us/minitab/help-and-how-to/statistical-modeling/anova/how-to/one-way-anova/before-you-start/data-considerations/
|
|
|
|
f_statistic, p_value = f_oneway(iso_group['tau_xfer'], ttx_group['tau_xfer'])
|
|
|
|
#%% f_crit control
|
|
|
|
dfn = 2-1 #2 gruppi, miinus 1, hetkel ei lahendanud seda vaga automatiseeritult
|
|
dfd = len(iso_group) + len(ttx_group) - 2 #ka ei lahendanud automatiseeritult
|
|
|
|
alpha = 0.05
|
|
|
|
f_crit = f.ppf(1 - alpha, dfn, dfd)
|
|
|
|
#%% Väljastame tulemused
|
|
|
|
print("F-statistika:", f_statistic)
|
|
print("P-väärtus:", p_value)
|
|
print("F-kriitiline:", f_crit)
|
|
"""
|