Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix - in create_dataset function, remove the unreachable return statement / remove extra spaces #5

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

fix - in create_dataset function, remove the unreachable return statement / remove extra spaces #5

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
191 changes: 85 additions & 106 deletions baselineUtils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,116 +9,103 @@

def create_dataset(dataset_folder,dataset_name,val_size,gt,horizon,delim="\t",train=True,eval=False,verbose=False):

if train==True:
datasets_list = os.listdir(os.path.join(dataset_folder,dataset_name, "train"))
full_dt_folder=os.path.join(dataset_folder,dataset_name, "train")
if train==False and eval==False:
datasets_list = os.listdir(os.path.join(dataset_folder, dataset_name, "val"))
full_dt_folder = os.path.join(dataset_folder, dataset_name, "val")
if train==False and eval==True:
datasets_list = os.listdir(os.path.join(dataset_folder, dataset_name, "test"))
full_dt_folder = os.path.join(dataset_folder, dataset_name, "test")


datasets_list=datasets_list
data={}
data_src=[]
data_trg=[]
data_seq_start=[]
data_frames=[]
data_dt=[]
data_peds=[]

val_src = []
val_trg = []
val_seq_start = []
val_frames = []
val_dt = []
val_peds=[]

if train==True:
datasets_list = os.listdir(os.path.join(dataset_folder,dataset_name, "train"))
full_dt_folder=os.path.join(dataset_folder,dataset_name, "train")
if train==False and eval==False:
datasets_list = os.listdir(os.path.join(dataset_folder, dataset_name, "val"))
full_dt_folder = os.path.join(dataset_folder, dataset_name, "val")
if train==False and eval==True:
datasets_list = os.listdir(os.path.join(dataset_folder, dataset_name, "test"))
full_dt_folder = os.path.join(dataset_folder, dataset_name, "test")


datasets_list=datasets_list
data={}
data_src=[]
data_trg=[]
data_seq_start=[]
data_frames=[]
data_dt=[]
data_peds=[]

val_src = []
val_trg = []
val_seq_start = []
val_frames = []
val_dt = []
val_peds=[]

if verbose:
print("start loading dataset")
print("validation set size -> %i"%(val_size))

for i_dt, dt in enumerate(datasets_list):
if verbose:
print("start loading dataset")
print("validation set size -> %i"%(val_size))


for i_dt, dt in enumerate(datasets_list):
if verbose:
print("%03i / %03i - loading %s"%(i_dt+1,len(datasets_list),dt))
raw_data = pd.read_csv(os.path.join(full_dt_folder, dt), delimiter=delim,
names=["frame", "ped", "x", "y"],usecols=[0,1,2,3],na_values="?")

raw_data.sort_values(by=['frame','ped'], inplace=True)

inp,out,info=get_strided_data_clust(raw_data,gt,horizon,1)

dt_frames=info['frames']
dt_seq_start=info['seq_start']
dt_dataset=np.array([i_dt]).repeat(inp.shape[0])
dt_peds=info['peds']



if val_size>0 and inp.shape[0]>val_size*2.5:
if verbose:
print("created validation from %s" % (dt))
k = random.sample(np.arange(inp.shape[0]).tolist(), val_size)
val_src.append(inp[k, :, :])
val_trg.append(out[k, :, :])
val_seq_start.append(dt_seq_start[k, :, :])
val_frames.append(dt_frames[k, :])
val_dt.append(dt_dataset[k])
val_peds.append(dt_peds[k])
inp = np.delete(inp, k, 0)
out = np.delete(out, k, 0)
dt_frames = np.delete(dt_frames, k, 0)
dt_seq_start = np.delete(dt_seq_start, k, 0)
dt_dataset = np.delete(dt_dataset, k, 0)
dt_peds = np.delete(dt_peds,k,0)
elif val_size>0:
if verbose:
print("could not create validation from %s, size -> %i" % (dt,inp.shape[0]))

data_src.append(inp)
data_trg.append(out)
data_seq_start.append(dt_seq_start)
data_frames.append(dt_frames)
data_dt.append(dt_dataset)
data_peds.append(dt_peds)


print("%03i / %03i - loading %s"%(i_dt+1,len(datasets_list),dt))
raw_data = pd.read_csv(os.path.join(full_dt_folder, dt), delimiter=delim,
names=["frame", "ped", "x", "y"],usecols=[0,1,2,3],na_values="?")

raw_data.sort_values(by=['frame','ped'], inplace=True)

inp,out,info=get_strided_data_clust(raw_data,gt,horizon,1)

data['src'] = np.concatenate(data_src, 0)
data['trg'] = np.concatenate(data_trg, 0)
data['seq_start'] = np.concatenate(data_seq_start, 0)
data['frames'] = np.concatenate(data_frames, 0)
data['dataset'] = np.concatenate(data_dt, 0)
data['peds'] = np.concatenate(data_peds, 0)
data['dataset_name'] = datasets_list
dt_frames=info['frames']
dt_seq_start=info['seq_start']
dt_dataset=np.array([i_dt]).repeat(inp.shape[0])
dt_peds=info['peds']

mean= data['src'].mean((0,1))
std= data['src'].std((0,1))

if val_size>0:
data_val={}
data_val['src']=np.concatenate(val_src,0)
data_val['trg'] = np.concatenate(val_trg, 0)
data_val['seq_start'] = np.concatenate(val_seq_start, 0)
data_val['frames'] = np.concatenate(val_frames, 0)
data_val['dataset'] = np.concatenate(val_dt, 0)
data_val['peds'] = np.concatenate(val_peds, 0)

return IndividualTfDataset(data, "train", mean, std), IndividualTfDataset(data_val, "validation", mean, std)
if val_size>0 and inp.shape[0]>val_size*2.5:
if verbose:
print("created validation from %s" % (dt))
k = random.sample(np.arange(inp.shape[0]).tolist(), val_size)
val_src.append(inp[k, :, :])
val_trg.append(out[k, :, :])
val_seq_start.append(dt_seq_start[k, :, :])
val_frames.append(dt_frames[k, :])
val_dt.append(dt_dataset[k])
val_peds.append(dt_peds[k])
inp = np.delete(inp, k, 0)
out = np.delete(out, k, 0)
dt_frames = np.delete(dt_frames, k, 0)
dt_seq_start = np.delete(dt_seq_start, k, 0)
dt_dataset = np.delete(dt_dataset, k, 0)
dt_peds = np.delete(dt_peds,k,0)
elif val_size>0:
if verbose:
print("could not create validation from %s, size -> %i" % (dt,inp.shape[0]))

return IndividualTfDataset(data, "train", mean, std), None
data_src.append(inp)
data_trg.append(out)
data_seq_start.append(dt_seq_start)
data_frames.append(dt_frames)
data_dt.append(dt_dataset)
data_peds.append(dt_peds)


data['src'] = np.concatenate(data_src, 0)
data['trg'] = np.concatenate(data_trg, 0)
data['seq_start'] = np.concatenate(data_seq_start, 0)
data['frames'] = np.concatenate(data_frames, 0)
data['dataset'] = np.concatenate(data_dt, 0)
data['peds'] = np.concatenate(data_peds, 0)
data['dataset_name'] = datasets_list

mean= data['src'].mean((0,1))
std= data['src'].std((0,1))

return IndividualTfDataset(data,"train",mean,std), IndividualTfDataset(data_val,"validation",mean,std)
if val_size>0:
data_val={}
data_val['src']=np.concatenate(val_src,0)
data_val['trg'] = np.concatenate(val_trg, 0)
data_val['seq_start'] = np.concatenate(val_seq_start, 0)
data_val['frames'] = np.concatenate(val_frames, 0)
data_val['dataset'] = np.concatenate(val_dt, 0)
data_val['peds'] = np.concatenate(val_peds, 0)

return IndividualTfDataset(data, "train", mean, std), IndividualTfDataset(data_val, "validation", mean, std)

return IndividualTfDataset(data, "train", mean, std), None

class IndividualTfDataset(Dataset):
def __init__(self,data,name,mean,std):
Expand All @@ -143,12 +130,6 @@ def __getitem__(self,index):
'peds': self.data['peds'][index],
}







def create_folders(baseFolder,datasetName):
try:
os.mkdir(baseFolder)
Expand All @@ -160,8 +141,6 @@ def create_folders(baseFolder,datasetName):
except:
pass



def get_strided_data(dt, gt_size, horizon, step):
inp_te = []
dtt = dt.astype(np.float32)
Expand Down