From ca08738c7b146238319aa229b786c650dfcce645 Mon Sep 17 00:00:00 2001
From: "@E181658" <celine.hourcade@etu.univ-nantes.fr>
Date: Wed, 21 Sep 2022 10:56:01 +0200
Subject: [PATCH] Creation of a single "discrim" function, csv file required
for all modes
---
README.md | 11 +-
data_process.py | 13 +-
output_demo/prediction_network_level.csv | 12 +-
output_demo/prediction_station_level.csv | 118 +++++++--------
prediction.py | 178 ++++++++---------------
run.py | 34 ++---
6 files changed, 148 insertions(+), 218 deletions(-)
diff --git a/README.md b/README.md
index f43d6c9..addcee4 100644
--- a/README.md
+++ b/README.md
@@ -40,6 +40,9 @@ To apply the algorithm, we need a folder architecture:
Each mseed file corresponds to the raw 3 component recordings of 60 sec.
+A csv file is also required to apply the algorithm.
+It is composed of a column with folders/events to discriminate. If the -valid option is specified, the file must have a second column with the label associated with the event.
+
## Trained model
Located in directory: model/model_2021354T1554.h5
@@ -47,11 +50,12 @@ Located in directory: model/model_2021354T1554.h5
## Prediction
+As input, you need a one-column csv file with the folders/events to be discriminated.
The algorithm supports the mseed data format.
- mseed format
```
- python run.py --mode pred --data_dir ./mseed_demo --output_dir ./output_demo
+ python run.py --data_dir ./mseed_demo --spectro_dir ./spectro_demo --output_dir ./output_demo --csv_dir demo_file.csv
```
Output files are automatically saved in "output_demo".
@@ -84,12 +88,13 @@ event and then average them. This gives us an event-based classification.
## Validation
-This mode can be used if the label is known. As input you need a csv file with the associated label for each event.
+This mode can be used if the label is known. The -valid argument is then specified.
+As input you need a csv file of two columns with the associated label for each event.
The algorithm supports the mseed data format.
- mseed format
```
- python run.py --mode valid --csv_dir ./demo_file.csv --data_dir ./mseed_demo --output_dir ./output_demo
+ python run.py --data_dir ./mseed_demo --spectro_dir ./spectro_demo --output_dir ./output_demo --csv_dir demo_file.csv --valid
```
Output files are automatically saved in "output_demo".
diff --git a/data_process.py b/data_process.py
index d1a5e2a..1365045 100644
--- a/data_process.py
+++ b/data_process.py
@@ -71,7 +71,7 @@ def get_fft(trace: op.core.trace.Trace, WINDOW_LENGTH: int,
def spectro_extract(data_dir: str, spectro_dir: str,
- events_list: list = None) -> None:
+ events_list: list) -> None:
"""
Compute the spectrograms that will be used for the validation.
The matrices are saved as NumPy objects.
@@ -87,10 +87,8 @@ def spectro_extract(data_dir: str, spectro_dir: str,
WINDOW_LENGTH = 1
OVERLAP = (1 - 0.75)
- if events_list is None:
- events = glob.glob(f'{data_dir}/*')
- else:
- events = events_list
+
+ events = events_list
print(f'Number of events: {len(events)}')
nb_evt = 0
@@ -99,10 +97,7 @@ def spectro_extract(data_dir: str, spectro_dir: str,
print('*****************')
print(f'EVENT {nb_evt} / {len(events)}')
- if events_list is None:
- time = events[a].split('/')[-1]
- else:
- time = events[a][0]
+ time = events[a][0]
if not os.path.exists(f'{spectro_dir}/{time}'):
os.makedirs(f'{spectro_dir}/{time}')
diff --git a/output_demo/prediction_network_level.csv b/output_demo/prediction_network_level.csv
index 776a646..05f7158 100644
--- a/output_demo/prediction_network_level.csv
+++ b/output_demo/prediction_network_level.csv
@@ -1,10 +1,10 @@
event,prob_nat,prob_ant,pred,nature
-2022004T111040,0.006,0.994,1,Anthropogenic
-2022001T213524,0.961,0.039,0,Natural
-2022003T080315,0.043,0.957,1,Anthropogenic
-2022003T084110,0.019,0.981,1,Anthropogenic
2022004T134407,0.77,0.23,0,Natural
+2022003T041502,0.988,0.012,0,Natural
+2022001T213524,0.961,0.039,0,Natural
+2022004T111745,0.005,0.995,1,Anthropogenic
+2022004T111040,0.006,0.994,1,Anthropogenic
2022004T105235,0.011,0.989,1,Anthropogenic
2022004T101445,0.015,0.985,1,Anthropogenic
-2022004T111745,0.005,0.995,1,Anthropogenic
-2022003T041502,0.988,0.012,0,Natural
+2022003T084110,0.019,0.981,1,Anthropogenic
+2022003T080315,0.043,0.957,1,Anthropogenic
diff --git a/output_demo/prediction_station_level.csv b/output_demo/prediction_station_level.csv
index 4a8b717..c99fc26 100644
--- a/output_demo/prediction_station_level.csv
+++ b/output_demo/prediction_station_level.csv
@@ -1,4 +1,39 @@
file_name,station,prob_nat,prob_ant,pred,nature
+FR_ABJF_2022004T134407,ABJF,0.044,0.956,1,Anthropogenic
+FR_CHLF_2022004T134407,CHLF,0.962,0.038,0,Natural
+FR_LBL_2022004T134407,LBL,0.216,0.784,1,Anthropogenic
+FR_GARF_2022004T134407,GARF,0.982,0.018,0,Natural
+FR_GNEF_2022004T134407,GNEF,0.914,0.086,0,Natural
+FR_VERF_2022004T134407,VERF,0.985,0.015,0,Natural
+FR_GZNF_2022004T134407,GZNF,0.985,0.015,0,Natural
+FR_COLF_2022004T134407,COLF,0.808,0.192,0,Natural
+FR_HRSF_2022004T134407,HRSF,0.989,0.011,0,Natural
+FR_BRGF_2022004T134407,BRGF,0.637,0.363,0,Natural
+FR_FRNF_2022004T134407,FRNF,0.952,0.048,0,Natural
+FR_PLDF_2022003T041502,PLDF,0.989,0.011,0,Natural
+FR_CHLF_2022003T041502,CHLF,0.985,0.015,0,Natural
+FR_AGO_2022003T041502,AGO,0.999,0.001,0,Natural
+FR_BRGF_2022003T041502,BRGF,0.988,0.012,0,Natural
+FR_FRNF_2022003T041502,FRNF,0.985,0.015,0,Natural
+FR_COLF_2022003T041502,COLF,0.986,0.014,0,Natural
+FR_HRSF_2022003T041502,HRSF,0.987,0.013,0,Natural
+FR_GZNF_2022003T041502,GZNF,0.997,0.003,0,Natural
+FR_VERF_2022003T041502,VERF,0.985,0.015,0,Natural
+FR_LBL_2022003T041502,LBL,0.977,0.023,0,Natural
+FR_MTNF_2022001T213524,MTNF,0.924,0.076,0,Natural
+FR_SROF_2022001T213524,SROF,0.961,0.039,0,Natural
+FR_PLOF_2022001T213524,PLOF,0.949,0.051,0,Natural
+FR_DAUF_2022001T213524,DAUF,0.996,0.004,0,Natural
+FR_PLEF_2022001T213524,PLEF,0.989,0.011,0,Natural
+FR_RIAF_2022001T213524,RIAF,0.999,0.001,0,Natural
+FR_LOCF_2022001T213524,LOCF,0.866,0.134,0,Natural
+FR_SOMF_2022001T213524,SOMF,0.944,0.056,0,Natural
+FR_LOUF_2022001T213524,LOUF,0.992,0.008,0,Natural
+FR_BESN_2022001T213524,BESN,0.991,0.009,0,Natural
+FR_PLEF_2022004T111745,PLEF,0.011,0.989,1,Anthropogenic
+FR_RIAF_2022004T111745,RIAF,0.006,0.994,1,Anthropogenic
+FR_BEGF_2022004T111745,BEGF,0.001,0.999,1,Anthropogenic
+FR_MTNF_2022004T111745,MTNF,0.001,0.999,1,Anthropogenic
FR_DAUF_2022004T111040,DAUF,0.001,0.999,1,Anthropogenic
FR_LEUC_2022004T111040,LEUC,0.002,0.998,1,Anthropogenic
FR_CRNF_2022004T111040,CRNF,0.014,0.986,1,Anthropogenic
@@ -14,51 +49,6 @@ FR_SLVF_2022004T111040,SLVF,0.001,0.999,1,Anthropogenic
FR_VERF_2022004T111040 2,VERF,0.014,0.986,1,Anthropogenic
FR_BRGF_2022004T111040,BRGF,0.005,0.995,1,Anthropogenic
FR_HRSF_2022004T111040,HRSF,0.003,0.997,1,Anthropogenic
-FR_MTNF_2022001T213524,MTNF,0.924,0.076,0,Natural
-FR_SROF_2022001T213524,SROF,0.961,0.039,0,Natural
-FR_PLOF_2022001T213524,PLOF,0.949,0.051,0,Natural
-FR_DAUF_2022001T213524,DAUF,0.996,0.004,0,Natural
-FR_PLEF_2022001T213524,PLEF,0.989,0.011,0,Natural
-FR_RIAF_2022001T213524,RIAF,0.999,0.001,0,Natural
-FR_LOCF_2022001T213524,LOCF,0.866,0.134,0,Natural
-FR_SOMF_2022001T213524,SOMF,0.944,0.056,0,Natural
-FR_LOUF_2022001T213524,LOUF,0.992,0.008,0,Natural
-FR_BESN_2022001T213524,BESN,0.991,0.009,0,Natural
-FR_CRNF_2022003T080315,CRNF,0.031,0.969,1,Anthropogenic
-FR_CHIF_2022003T080315 2,CHIF,0.047,0.953,1,Anthropogenic
-FR_GNEF_2022003T080315 2,GNEF,0.014,0.986,1,Anthropogenic
-FR_GARF_2022003T080315 2,GARF,0.008,0.992,1,Anthropogenic
-FR_LGIF_2022003T080315 2,LGIF,0.014,0.986,1,Anthropogenic
-FR_CHIF_2022003T080315,CHIF,0.047,0.953,1,Anthropogenic
-FR_BRGF_2022003T080315 2,BRGF,0.162,0.838,1,Anthropogenic
-FR_ABJF_2022003T080315 2,ABJF,0.04,0.96,1,Anthropogenic
-FR_CRNF_2022003T080315 2,CRNF,0.031,0.969,1,Anthropogenic
-FR_ABJF_2022003T080315,ABJF,0.04,0.96,1,Anthropogenic
-FR_BSCF_2022003T080315 2,BSCF,0.03,0.97,1,Anthropogenic
-FR_GNEF_2022003T080315,GNEF,0.014,0.986,1,Anthropogenic
-FR_GARF_2022003T080315,GARF,0.008,0.992,1,Anthropogenic
-FR_LGIF_2022003T080315,LGIF,0.014,0.986,1,Anthropogenic
-FR_BSCF_2022003T080315,BSCF,0.03,0.97,1,Anthropogenic
-FR_BRGF_2022003T080315,BRGF,0.162,0.838,1,Anthropogenic
-FR_CHIF_2022003T084110,CHIF,0.006,0.994,1,Anthropogenic
-FR_RIAF_2022003T084110,RIAF,0.039,0.961,1,Anthropogenic
-FR_CRNF_2022003T084110,CRNF,0.014,0.986,1,Anthropogenic
-FR_BOUF_2022003T084110,BOUF,0.054,0.946,1,Anthropogenic
-FR_DAUF_2022003T084110,DAUF,0.011,0.989,1,Anthropogenic
-FR_GNEF_2022003T084110,GNEF,0.011,0.989,1,Anthropogenic
-FR_LGIF_2022003T084110,LGIF,0.012,0.988,1,Anthropogenic
-FR_BSCF_2022003T084110,BSCF,0.004,0.996,1,Anthropogenic
-FR_ABJF_2022004T134407,ABJF,0.044,0.956,1,Anthropogenic
-FR_CHLF_2022004T134407,CHLF,0.962,0.038,0,Natural
-FR_LBL_2022004T134407,LBL,0.216,0.784,1,Anthropogenic
-FR_GARF_2022004T134407,GARF,0.982,0.018,0,Natural
-FR_GNEF_2022004T134407,GNEF,0.914,0.086,0,Natural
-FR_VERF_2022004T134407,VERF,0.985,0.015,0,Natural
-FR_GZNF_2022004T134407,GZNF,0.985,0.015,0,Natural
-FR_COLF_2022004T134407,COLF,0.808,0.192,0,Natural
-FR_HRSF_2022004T134407,HRSF,0.989,0.011,0,Natural
-FR_BRGF_2022004T134407,BRGF,0.637,0.363,0,Natural
-FR_FRNF_2022004T134407,FRNF,0.952,0.048,0,Natural
FR_DAUF_2022004T105235,DAUF,0.003,0.997,1,Anthropogenic
FR_BEGF_2022004T105235,BEGF,0.009,0.991,1,Anthropogenic
FR_BOUF_2022004T105235,BOUF,0.022,0.978,1,Anthropogenic
@@ -78,17 +68,27 @@ FR_SROF_2022004T101445,SROF,0.061,0.939,1,Anthropogenic
FR_MTNF_2022004T101445,MTNF,0.006,0.994,1,Anthropogenic
FR_GUEF_2022004T101445,GUEF,0.002,0.998,1,Anthropogenic
FR_OLIV_2022004T101445,OLIV,0.004,0.996,1,Anthropogenic
-FR_PLEF_2022004T111745,PLEF,0.011,0.989,1,Anthropogenic
-FR_RIAF_2022004T111745,RIAF,0.006,0.994,1,Anthropogenic
-FR_BEGF_2022004T111745,BEGF,0.001,0.999,1,Anthropogenic
-FR_MTNF_2022004T111745,MTNF,0.001,0.999,1,Anthropogenic
-FR_PLDF_2022003T041502,PLDF,0.989,0.011,0,Natural
-FR_CHLF_2022003T041502,CHLF,0.985,0.015,0,Natural
-FR_AGO_2022003T041502,AGO,0.999,0.001,0,Natural
-FR_BRGF_2022003T041502,BRGF,0.988,0.012,0,Natural
-FR_FRNF_2022003T041502,FRNF,0.985,0.015,0,Natural
-FR_COLF_2022003T041502,COLF,0.986,0.014,0,Natural
-FR_HRSF_2022003T041502,HRSF,0.987,0.013,0,Natural
-FR_GZNF_2022003T041502,GZNF,0.997,0.003,0,Natural
-FR_VERF_2022003T041502,VERF,0.985,0.015,0,Natural
-FR_LBL_2022003T041502,LBL,0.977,0.023,0,Natural
+FR_CHIF_2022003T084110,CHIF,0.006,0.994,1,Anthropogenic
+FR_RIAF_2022003T084110,RIAF,0.039,0.961,1,Anthropogenic
+FR_CRNF_2022003T084110,CRNF,0.014,0.986,1,Anthropogenic
+FR_BOUF_2022003T084110,BOUF,0.054,0.946,1,Anthropogenic
+FR_DAUF_2022003T084110,DAUF,0.011,0.989,1,Anthropogenic
+FR_GNEF_2022003T084110,GNEF,0.011,0.989,1,Anthropogenic
+FR_LGIF_2022003T084110,LGIF,0.012,0.988,1,Anthropogenic
+FR_BSCF_2022003T084110,BSCF,0.004,0.996,1,Anthropogenic
+FR_CRNF_2022003T080315,CRNF,0.031,0.969,1,Anthropogenic
+FR_CHIF_2022003T080315 2,CHIF,0.047,0.953,1,Anthropogenic
+FR_GNEF_2022003T080315 2,GNEF,0.014,0.986,1,Anthropogenic
+FR_GARF_2022003T080315 2,GARF,0.008,0.992,1,Anthropogenic
+FR_LGIF_2022003T080315 2,LGIF,0.014,0.986,1,Anthropogenic
+FR_CHIF_2022003T080315,CHIF,0.047,0.953,1,Anthropogenic
+FR_BRGF_2022003T080315 2,BRGF,0.162,0.838,1,Anthropogenic
+FR_ABJF_2022003T080315 2,ABJF,0.04,0.96,1,Anthropogenic
+FR_CRNF_2022003T080315 2,CRNF,0.031,0.969,1,Anthropogenic
+FR_ABJF_2022003T080315,ABJF,0.04,0.96,1,Anthropogenic
+FR_BSCF_2022003T080315 2,BSCF,0.03,0.97,1,Anthropogenic
+FR_GNEF_2022003T080315,GNEF,0.014,0.986,1,Anthropogenic
+FR_GARF_2022003T080315,GARF,0.008,0.992,1,Anthropogenic
+FR_LGIF_2022003T080315,LGIF,0.014,0.986,1,Anthropogenic
+FR_BSCF_2022003T080315,BSCF,0.03,0.97,1,Anthropogenic
+FR_BRGF_2022003T080315,BRGF,0.162,0.838,1,Anthropogenic
diff --git a/prediction.py b/prediction.py
index eb0dfbc..ae17472 100644
--- a/prediction.py
+++ b/prediction.py
@@ -10,7 +10,7 @@ from numpy import moveaxis
import tensorflow as tf
-def pred(model_dir, spectro_dir, output_dir):
+def discrim(spectro_dir, output_dir, event_label, valid):
"""
Event class prediction.
@@ -20,140 +20,63 @@ def pred(model_dir, spectro_dir, output_dir):
:param spectro_dir: Absolute path to the input spectrograms.
:type output_dir: str
:param output_dir: Absolute path where to save to output files.
+ :type event_label: list
+ :param event_label: The class of event to validate.
"""
-
- csvPr_sta = open(os.path.join(
- output_dir, 'prediction_station_level.csv'), 'w')
- predict_sta = csv.writer(csvPr_sta, delimiter=',',
- quotechar='"', quoting=csv.QUOTE_MINIMAL)
- predict_sta.writerow(['file_name',
+
+ if valid :
+ filename_csvsta = 'validation_station_level.csv'
+ csvsta_row = ['file_name',
'station',
+ 'label_cat',
'prob_nat',
'prob_ant',
'pred',
- 'nature',
- ])
+ 'nature']
- csvPr_net = open(os.path.join(
- output_dir, 'prediction_network_level.csv'), 'w')
- predict_net = csv.writer(csvPr_net, delimiter=',',
- quotechar='"', quoting=csv.QUOTE_MINIMAL)
- predict_net.writerow(['event',
+ filename_csvnet = 'validation_network_level.csv'
+ csvnet_row = ['event',
+ 'label_cat',
'prob_nat',
'prob_ant',
'pred',
- 'nature',
- ])
-
- model = tf.keras.models.load_model(model_dir)
-
- events = glob.glob(f'{spectro_dir}/*')
-
- print(f'Number of events: {len(events)}')
-
- nb_evt = 0
- for evt in events:
- nb_evt += 1
- print('*****************')
- print(f'EVENT {nb_evt} / {len(events)}')
-
- time = evt.split('/')[-1]
- pred_nat = 0
- pred_ant = 0
-
- list_spect = glob.glob(f'{spectro_dir}/{time}/*')
- print(f'Number of station: {len(list_spect)}')
- nb_st = 0
- for spect in list_spect:
- nb_st += 1
- print(f'Station {nb_st} / {len(list_spect)}', end = "\r")
- file_name = (spect.split('/')[-1]).split('.npy')[0]
-
- station = file_name.split('_')[1]
- spect_file = np.load(f'{spect}', allow_pickle=True)
- spect_file = [np.array(spect_file)]
-
- x = moveaxis(spect_file, 1, 3)
-
- model_output = model.predict(x).round(3)
- pred = np.argmax(model_output, axis=1)
-
- if pred == 0:
- pred_final = 'Natural'
- if pred == 1:
- pred_final = 'Anthropogenic'
-
- predict_sta.writerow([file_name,
- station,
- model_output[0][0],
- model_output[0][1],
- pred[0],
- pred_final,
- ])
+ 'nature']
- pred_nat += model_output[0][0]
- pred_ant += model_output[0][1]
-
- pred_total = [pred_nat, pred_ant]
- pred_total = [(float(i)/sum(pred_total)).round(3) for i in pred_total]
- pred_event = np.argmax(pred_total)
- if pred_event == 0:
- pred_final = 'Natural'
- if pred_event == 1:
- pred_final = 'Anthropogenic'
-
- predict_net.writerow([time,
- pred_total[0],
- pred_total[1],
- pred_event,
- pred_final,
- ])
-
-
-def valid(model_dir, spectro_dir, output_dir, event_label):
- """
- Event class validation.
-
- :type model_dir: str
- :param model_dir: Absolute path the the trained model.
- :type spectro_dir: str
- :param spectro_dir: Absolute path to the input spectrograms.
- :type output_dir: str
- :param output_dir: Absolute path where to save the output files.
- :type event_label: list
- :param event_label: The class of event to validate.
- """
-
- csvPr_sta = open(os.path.join(
- output_dir, 'validation_station_level.csv'), 'w')
- predict_sta = csv.writer(csvPr_sta, delimiter=',',
- quotechar='"', quoting=csv.QUOTE_MINIMAL)
- predict_sta.writerow(['file_name',
+ else :
+ filename_csvsta = 'prediction_station_level.csv'
+ csvsta_row = ['file_name',
'station',
- 'label_cat',
'prob_nat',
'prob_ant',
'pred',
- 'nature',
- ])
+ 'nature']
- csvPr_net = open(os.path.join(
- output_dir, 'validation_network_level.csv'), 'w')
- predict_net = csv.writer(csvPr_net, delimiter=',',
- quotechar='"', quoting=csv.QUOTE_MINIMAL)
- predict_net.writerow(['event',
- 'label_cat',
+ filename_csvnet = 'prediction_network_level.csv'
+ csvnet_row = ['event',
'prob_nat',
'prob_ant',
'pred',
- 'nature'])
+ 'nature'
+ ]
+
+ csvPr_sta = open(os.path.join(
+ output_dir, filename_csvsta), 'w')
+ predict_sta = csv.writer(csvPr_sta, delimiter=',',
+ quotechar='"', quoting=csv.QUOTE_MINIMAL)
+ predict_sta.writerow(csvsta_row)
- model = tf.keras.models.load_model(model_dir)
+ csvPr_net = open(os.path.join(
+ output_dir, filename_csvnet), 'w')
+ predict_net = csv.writer(csvPr_net, delimiter=',',
+ quotechar='"', quoting=csv.QUOTE_MINIMAL)
+ predict_net.writerow(csvnet_row)
+
+ model = tf.keras.models.load_model("./model/model_2021354T1554.h5")
events = glob.glob(f'{spectro_dir}/*')
print(f'Number of events: {len(events)}')
- print(len(event_label))
+
nb_evt = 0
for a in range(len(event_label)):
nb_evt += 1
@@ -161,8 +84,8 @@ def valid(model_dir, spectro_dir, output_dir, event_label):
print(f'EVENT {nb_evt} / {len(event_label)}')
time = event_label[a][0]
- class_ = event_label[a][1]
- #time = evt.split('/')[2]
+ if valid:
+ class_ = event_label[a][1]
pred_nat = 0
pred_ant = 0
@@ -187,8 +110,9 @@ def valid(model_dir, spectro_dir, output_dir, event_label):
pred_final = 'Natural'
if pred == 1:
pred_final = 'Anthropogenic'
-
- predict_sta.writerow([file_name,
+
+ if valid :
+ predict_sta.writerow([file_name,
station,
class_,
model_output[0][0],
@@ -196,6 +120,14 @@ def valid(model_dir, spectro_dir, output_dir, event_label):
pred[0],
pred_final,
])
+ else :
+ predict_sta.writerow([file_name,
+ station,
+ model_output[0][0],
+ model_output[0][1],
+ pred[0],
+ pred_final,
+ ])
pred_nat += model_output[0][0]
pred_ant += model_output[0][1]
@@ -207,11 +139,21 @@ def valid(model_dir, spectro_dir, output_dir, event_label):
pred_final = 'Natural'
if pred_event == 1:
pred_final = 'Anthropogenic'
-
- predict_net.writerow([time,
+
+ if valid :
+ predict_net.writerow([time,
class_,
pred_total[0],
pred_total[1],
pred_event,
pred_final,
])
+ else :
+ predict_net.writerow([time,
+ pred_total[0],
+ pred_total[1],
+ pred_event,
+ pred_final,
+ ])
+
+
diff --git a/run.py b/run.py
index 41090c5..cd4a2b5 100644
--- a/run.py
+++ b/run.py
@@ -5,7 +5,7 @@ import argparse
import numpy as np
-from prediction import pred, valid
+from prediction import discrim
from data_process import spectro_extract
@@ -13,14 +13,6 @@ def read_args() -> argparse.Namespace:
parser = argparse.ArgumentParser()
- parser.add_argument('--mode',
- type=str, default='pred',
- help=' "valid" for the validation mode or "pred" for the prediction mode.')
-
- parser.add_argument('--model_dir',
- type=str, default='./model/model_2021354T1554.h5',
- help="Model file directory.")
-
parser.add_argument('--data_dir',
type=str, default='./mseed_demo',
help="Input mseed file directory.")
@@ -30,36 +22,32 @@ def read_args() -> argparse.Namespace:
help='Output spectrogram file directory.')
parser.add_argument('--csv_dir',
- default=None,
+ required=True,
help="Input csv file directory")
parser.add_argument('--output_dir',
type=str, default='./output_demo',
help='Output directory')
+ parser.add_argument('--valid',
+ action="store_true",
+ help=' if the option "valid" is specified the validation mode will be applied. Csv input must have two columns (time, label_cat)')
+
args = parser.parse_args()
return args
def main(args: argparse.Namespace):
- if args.mode == 'pred':
- spectro_extract(data_dir=args.data_dir,
- spectro_dir=args.spectro_dir)
- pred(model_dir=args.model_dir, spectro_dir=args.spectro_dir,
- output_dir=args.output_dir)
- elif args.mode == 'valid':
- events = np.genfromtxt(
+ events = np.genfromtxt(
f'{args.csv_dir}', delimiter=',', skip_header=1, dtype=str)
- spectro_extract(data_dir=args.data_dir,
+ spectro_extract(data_dir=args.data_dir,
spectro_dir=args.spectro_dir, events_list=events)
- valid(model_dir=args.model_dir, spectro_dir=args.spectro_dir,
- output_dir=args.output_dir, event_label=events)
-
- else:
- print("Mode should be: valid, or pred")
+ discrim(spectro_dir=args.spectro_dir,
+ output_dir=args.output_dir, event_label=events, valid = args.valid)
+
return
--
GitLab