def test_load():
data = [1, 2, 3, 4, 5]
assert pypickle.save('test.pkl', data, overwrite=True)
data_load = pypickle.load('test.pkl')
assert data_load == data
def process_csv_file(uploaded_filenames, uploaded_file_contents, y_min, alpha, k, excl_background, perc_min_num, specificity, multtest):
"""Save uploaded files and regenerate the file list."""
# Check input parameters
[args, runOK, runtxt]=check_input(uploaded_filenames, uploaded_file_contents, y_min, alpha, k, excl_background, perc_min_num, specificity, multtest)
if runOK==False:
for txt in runtxt: print('[HNET-GUI] %s' %txt)
return(runtxt)
print('alpha:%s' %args['alpha'])
print('y_min:%s' %args['y_min'])
print('k:%s' %args['k'])
print('multtest:%s' %args['multtest'])
print('excl_background:%s' %args['excl_background'])
print('perc_min_num:%s' %args['perc_min_num'])
print('specificity:%s' %args['specificity'])
print('dropna:%s' %args['dropna'])
print('File input: %s' %(args['uploaded_filenames']))
# if uploaded_filenames is not None and uploaded_file_contents is not None:
filepath = save_file(args['uploaded_filenames'], args['uploaded_file_contents'], TMP_DIRECTORY)
[_,filename, _] = hnet.hnet.path_split(args['uploaded_filenames'])
savepath = os.path.join(HNET_DIR_STABLE,filename+'_'+str(args['y_min'])+'_'+str(args['k'])+'_'+str(args['multtest'])+'_'+str(args['specificity'])+'_'+str(args['perc_min_num'])+'_'+str(args['excl_background'])+'/')
d3path = get_d3path(savepath)
pklpath = get_pklpath(savepath)
print('filepath %s' %(filepath))
print('savepath %s' %(savepath))
print('d3path %s' %(d3path))
print('pklpath %s' %(pklpath))
# Progressbar
# mGUI=tk.Tk()
# mGUI.title('HNET '+ filename)
# progressbar = ttk.Progressbar(mGUI, orient='horizontal', length=300)
# progressbar.pack(side=tk.TOP)
# progressbar.config(mode='determinate')
# progressbar.update()
# progressbar['value']=0
# progressbar['maximum']=5
# # Set starting position
# progressbar['value']=progressbar['value']+1 # update bar
# progressbar.update() # update gui
# Make directory
if not os.path.isdir(savepath):
os.mkdir(savepath)
# Make D3js path
if not os.path.isfile(d3path):
# Read file
df = pd.read_csv(filepath)
# Run HNet
# progressbar['value']=progressbar['value']+1 # update bar
# progressbar.update() # update gui
HNET_OUT = hnet.fit(df, alpha=args['alpha'], y_min=args['y_min'], k=args['k'], multtest=args['multtest'], dtypes='pandas', specificity=args['specificity'], perc_min_num=args['perc_min_num'], dropna=args['dropna'], excl_background=args['excl_background'], verbose=3)
# Save pickle file
print('SAVING NETWORK FIGURE: %s' %(savepath))
# progressbar['value']=progressbar['value']+1 # update bar
# progressbar.update() # update gui
HNET_OUT['G'] = hnet.plot_network(HNET_OUT, dist_between_nodes=0.4, scale=2, dpi=250, figsize=[30,20], showfig=False, savepath=os.path.join(savepath,'hnet_network.png'))
# Store pickle file
print('STORE PICKLE')
# progressbar['value']=progressbar['value']+1 # update bar
# progressbar.update() # update gui
pypickle.save(pklpath, HNET_OUT)
#print('MAKE D3GRAPH')
_ = hnet.plot_d3graph(HNET_OUT, savepath=savepath, showfig=False)
else:
print('dir exists, load stuff')
HNET_OUT=pypickle.load(pklpath)
# Open in browser
if os.path.isfile(d3path):
print('OPEN BROWSER')
webbrowser.open(os.path.abspath(d3path), new=2)
print('HNET_DIR_STABLE: %s!' %(HNET_DIR_STABLE))
print('%s done!' %(filename))
print('-----------------------Done!-----------------------')
# progressbar['value']=progressbar['maximum'] # update bar
# progressbar.update() # update gui
# try:
# mGUI.destroy()
# except:
# pass
return(filename)
def test_save():
data = [1, 2, 3, 4, 5]
assert pypickle.save('test.pkl', data, overwrite=True)
assert pypickle.save('test.pkl', data, overwrite=False) == False
# %% import pypickle print(dir(pypickle)) print(pypickle.__version__) # %% filepath = 'tes1t.pkl' data = [1, 2, 3, 4, 5] status = pypickle.save(filepath, data, fix_imports=True, overwrite=True) #%% Load file data = pypickle.load(filepath, encoding="latin1") # %%
def save_proxy_list(proxy_list):
proxy_list_name = get_proxy_list_file_name()
pypickle.save(proxy_list, proxy_list_name)
def save_proxy_page(page):
file = get_proxy_page_file_name()
pypickle.save(page, file)
out['HNETGradientScoresDirected']=HNETGradientScoresDirected
out['HNETGradientScoresUnDirected']=HNETGradientScoresUnDirected
out['HNETGradientScores']=HNETGradientScores
out['df_method_comparison']=df_method_comparison
out['df_method_comparison_scores']=df_method_comparison_scores
# Random scores
out['scores_rand']=scores_rand
#Best performing results
idx=HNETGradientScores['mcc_undirected'].argmax()
out['hnet_undirected_bestmodel_args']=HNETGradientScores.iloc[idx,:]
out['hnet_directed_bestmodel_args']=HNETGradientScores.iloc[idx,:]
out['hnet_undirected_bestmodel']=out_hnets['adjmat'][out_hnets['adjmat']>HNETGradientScores.iloc[idx,:]['minP']].fillna(0)
#%% Save
savename='hnet_'+arg['DAG']+'_'+str(arg['n_sampling'])+'.pkl'
pypickle.save(os.path.join('../PROJECTS/hnet/results/',savename), out)
#%%
print(arg)
print(df_method_comparison)
print(df_method_comparison_scores)
#%% Plot all results with these arguments
# Real model
_ = bayes.plot(modelTrue, pos=modelTrue['G']['pos'])
## Learned model using bayes
_=bayes.plot(out_bayes['adjmat'], pos=modelTrue['G']['pos'])
## Learned model using hnets
_=bayes.plot(out_hnets['adjmat']>0, pos=modelTrue['G']['pos'])
#%% MAKE FIGURE
def create_labels(url=path):
aliment_ids = {}
with open(url + "img_annotations.json", "r") as json_file:
photos = json.load(json_file)
for k, v in photos.items():
aliment_ids.setdefault(k, [])
aliment_ids[k] = [x["id"] for x in v]
return aliment_ids
def get_labels(tomatoes=get_ids_tomatoes(), aliment_ids=create_labels()):
labels = {}
for k, v in aliment_ids.items():
labels.setdefault(k, )
labels[k] = 0
for tomato in tomatoes:
if tomato in v:
labels[k] = 1
break
return labels
if __name__ == '__main__':
print(get_labels())
pickle.save(var=get_labels(),
filepath='/Users/RomainLejeune/Downloads/tomato_label.pkl')