Forecast Scheduling - WellsGroup Classes¶
WellsGroup class in the one on top of the others. You can specify a group of Well instances that, as seen before, each of them can be groups of scenarios and periods.
WellsGroup has different functionality when evaluating it. Each Well part of this have n scenarios which are independent to each other, so if you would like to evaluate all scenarios in all wells you would have to sample one scenario per well and build all the possible combinations. To do this, Dcapy uses the functions of PyDOE2 to generate factorial designs either full-factorial (fullfact) or Generalized Subset Designs (gds).
On the other hand, in this notebook is applied another convinient way of creating a dca model, which is by providing a yml file.
In [6]:
Copied!
import pandas as pd
import numpy as np
from dcapy import dca
from dcapy.schedule import Well, Period, Scenario, WellsGroup, model_from_dict
from dcapy.cashflow import CashFlowParams, CashFlow
from dcapy.wiener import Brownian, GeometricBrownian, MeanReversion
from dcapy.auth import Credential
import seaborn as sns
from datetime import date
import matplotlib.pyplot as plt
import copy
import yaml
import json
from scipy import stats
import pandas as pd
import numpy as np
from dcapy import dca
from dcapy.schedule import Well, Period, Scenario, WellsGroup, model_from_dict
from dcapy.cashflow import CashFlowParams, CashFlow
from dcapy.wiener import Brownian, GeometricBrownian, MeanReversion
from dcapy.auth import Credential
import seaborn as sns
from datetime import date
import matplotlib.pyplot as plt
import copy
import yaml
import json
from scipy import stats
In [7]:
Copied!
cred = Credential(token='eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ImIyZDQ5NjMyLWM0MzEtNDAzYi04OTEyLTJiZGIyOTA3NTMxNCIsIm5hbWUiOiJTYW50aWFnbyIsImxhc3RfbmFtZSI6IkN1ZXJ2byIsInVzZXJuYW1lIjoic2N1ZXJ2bzkxIiwiZXhwIjoxNjI2OTI2NTk3fQ.n3HuheJvoQKF9RNKTC9gEstC449EWd2qsrWR7f30V2U')
cred = Credential(token='eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ImIyZDQ5NjMyLWM0MzEtNDAzYi04OTEyLTJiZGIyOTA3NTMxNCIsIm5hbWUiOiJTYW50aWFnbyIsImxhc3RfbmFtZSI6IkN1ZXJ2byIsInVzZXJuYW1lIjoic2N1ZXJ2bzkxIiwiZXhwIjoxNjI2OTI2NTk3fQ.n3HuheJvoQKF9RNKTC9gEstC449EWd2qsrWR7f30V2U')
In a file called YML_example1.yml is a WellsGroup case where are defined 6 wells. Four of them have one scenario and others two have two scenarios with two Periods.
name: fdp_field
wells:
DC2:
name: DC2
scenarios:
base:
name: base
periods:
pdp:
dca:
bsw: 0.9783
fluid_rate: 5192.343117182031
slope: 3.986053847389352e-05
ti: '2021-04-01'
end: '2030-12-31'
freq_input: D
freq_output: D
name: pdp
start: '2021-04-01'
rate_limit: 50
cashflow_params:
- name: capex_abandon
periods: -1
target: capex
value: -200000.0
DC3:
name: DC3
scenarios:
base:
name: base
periods:
pdp:
dca:
bsw: 0.8935
fluid_rate: 3101.6068269182692
slope: 1.1341265851141962e-05
ti: '2021-04-01'
end: '2030-12-31'
freq_input: D
freq_output: D
name: pdp
start: '2021-04-01'
rate_limit: 50
cashflow_params:
- name: capex_abandon
periods: -1
target: capex
value: -200000.0
DC4:
name: DC4
scenarios:
base:
name: base
periods:
pdp:
dca:
bsw: 0.9863
fluid_rate: 7304.560639994402
slope: 4.086888672335466e-05
ti: '2021-04-01'
end: '2030-12-31'
freq_input: D
freq_output: D
name: pdp
start: '2021-04-01'
rate_limit: 50
cashflow_params:
- name: capex_abandon
periods: -1
target: capex
value: -200000.0
DC5:
name: DC5
scenarios:
base:
name: base
periods:
pdp:
dca:
bsw: 0.9648
fluid_rate: 5710.712047244095
slope: 1.9260822570459275e-05
ti: '2021-04-01'
end: '2030-12-31'
freq_input: D
freq_output: D
name: pdp
start: '2020-04-01'
rate_limit: 50
cashflow_params:
- name: capex_abandon
periods: -1
target: capex
value: -200000.0
well-1:
name: well-1
scenarios:
highfr:
name: highfr
periods:
fm1:
cashflow_params:
- name: capex
periods: 1
target: capex
value: -4000000.0
dca:
bsw:
dist: uniform
kw:
loc: 0.4
scale: 0.4
fluid_rate: 6500.0
slope:
- 3.0e-05
ti: '2022-01-01'
end: '2027-01-01'
name: fm1
rate_limit: 200.0
start: '2022-01-01'
fm2:
cashflow_params:
- name: capex_wo
periods: 1
target: capex
value: -600000.0
- name: capex_abandon
periods: -1
target: capex
value: -200000.0
dca:
bsw:
dist: uniform
kw:
loc: 0.4
scale: 0.4
fluid_rate: 6500.0
slope:
- 5.0e-05
ti: '2022-01-01'
depends:
period: fm1
end: '2030-01-01'
name: fm2
start: '2022-01-01'
mediumfr:
name: mediumfr
periods:
fm1:
cashflow_params:
- name: capex_drill
periods: 1
target: capex
value: -4000000.0
dca:
bsw:
dist: uniform
kw:
loc: 0.4
scale: 0.4
fluid_rate: 3500.0
slope:
- 3.0e-06
ti: '2022-01-01'
end: '2030-01-01'
name: fm1
rate_limit: 200.0
start: '2022-01-01'
fm2:
cashflow_params:
- name: capex_wo
periods: 1
target: capex
value: -600000.0
- name: capex_abandon
periods: -1
target: capex
value: -200000.0
dca:
bsw:
dist: uniform
kw:
loc: 0.4
scale: 0.4
fluid_rate: 2500.0
slope:
- 5.0e-06
ti: '2022-01-01'
depends:
period: fm1
end: '2030-01-01'
name: fm2
start: '2022-01-01'
well-2:
name: well-2
scenarios:
highfr:
name: highfr
periods:
fm2:
cashflow_params:
- name: capex_drill
periods: 1
target: capex
value: -4000000.0
dca:
bsw:
dist: uniform
kw:
loc: 0.4
scale: 0.4
fluid_rate: 6500.0
slope:
- 5.0e-05
ti: '2022-01-01'
end: '2030-01-01'
name: fm2
rate_limit: 200.0
start: '2022-01-01'
fm1:
cashflow_params:
- name: capex_wo
periods: 1
target: capex
value: -600000.0
- name: capex_abandon
periods: -1
target: capex
value: -200000.0
dca:
bsw:
dist: uniform
kw:
loc: 0.4
scale: 0.4
fluid_rate: 6500.0
slope:
- 3.0e-05
ti: '2022-01-01'
depends:
period: fm2
end: '2030-01-01'
name: fm1
start: '2022-01-01'
mediumfr:
name: mediumfr
periods:
fm2:
cashflow_params:
- name: capex_drill
periods: 1
target: capex
value: -4000000.0
dca:
bsw:
dist: uniform
kw:
loc: 0.4
scale: 0.4
fluid_rate: 2500.0
slope:
- 5.0e-06
ti: '2022-01-01'
end: '2030-01-01'
name: fm2
rate_limit: 200.0
start: '2022-01-01'
fm1:
cashflow_params:
- name: capex_wo
periods: 1
target: capex
value: -600000.0
- name: capex_abandon
periods: -1
target: capex
value: -200000.0
dca:
bsw:
dist: uniform
kw:
loc: 0.4
scale: 0.4
fluid_rate: 3500.0
slope:
- 3.0e-06
ti: '2022-01-01'
depends:
period: fm2
end: '2030-01-01'
name: fm1
start: '2022-01-01'
cashflow_params:
- name: fix_opex
target: opex
value: -11000.
freq_value: 'M'
- iter: 1
multiply: oil_volume
name: var_opex
target: opex
value: -8.0
- multiply: oil_volume
name: income
target: income
value:
initial_condition: 60
ti: '2021-04-01'
steps: 11
generator:
dist: norm
kw:
loc: 0.
scale: 13.13
m: 46.77
eta: 0.112653
freq_input: A
wi: 0.92
- name: buy
periods: 1
target: capex
value: -15500000
general: True
seed: 21
By using a YML parser like PyYAML you can convery a *.yml file into a python dictionary.
In [8]:
Copied!
with open('YML_example1.yml','r') as file:
case_dict = yaml.load(file)
with open('YML_example1.yml','r') as file:
case_dict = yaml.load(file)
<ipython-input-8-33d4e64ac547>:2: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details. case_dict = yaml.load(file)
The resulting dictionary, if it has valid values for key arguments, can be passed to a WellsGroup directly as made before on the others Schedule Classes
In [9]:
Copied!
case = WellsGroup(**case_dict)
print(type(case))
case = WellsGroup(**case_dict)
print(type(case))
<class 'dcapy.schedule.schedule.WellsGroup'>
WellsGroup has a method to make the all the factorial combinations of wells scenarios. By calling it with no arguments it creates a full-factorial combinations.
In [10]:
Copied!
case.insert_db(cred, description='Wellsgroup2')
case.insert_db(cred, description='Wellsgroup2')
Out[10]:
'5f625d24-517d-4890-8067-f8e4da41f779'
In [11]:
Copied!
case.update_db(cred, description='Tutorial-Wellsgroup')
case.update_db(cred, description='Tutorial-Wellsgroup')
Out[11]:
'5f625d24-517d-4890-8067-f8e4da41f779'
In [13]:
Copied!
cased = WellsGroup()
cased.get_db("5f625d24-517d-4890-8067-f8e4da41f779",cred)
cased = WellsGroup()
cased.get_db("5f625d24-517d-4890-8067-f8e4da41f779",cred)
In [15]:
Copied!
type(cased)
type(cased)
Out[15]:
dcapy.schedule.schedule.WellsGroup
Get the tree Schema¶
In [16]:
Copied!
case.tree()
case.tree()
Out[16]:
🏭fdp_field ┣━━ 🗼DC2 ┃ ┗━━ 🔀base ┃ ┗━━ 📉pdp ┣━━ 🗼DC3 ┃ ┗━━ 🔀base ┃ ┗━━ 📉pdp ┣━━ 🗼DC4 ┃ ┗━━ 🔀base ┃ ┗━━ 📉pdp ┣━━ 🗼DC5 ┃ ┗━━ 🔀base ┃ ┗━━ 📉pdp ┣━━ 🗼well-1 ┃ ┣━━ 🔀highfr ┃ ┃ ┣━━ 📉fm1 ┃ ┃ ┗━━ 📉fm2 ┃ ┗━━ 🔀mediumfr ┃ ┣━━ 📉fm1 ┃ ┗━━ 📉fm2 ┗━━ 🗼well-2 ┣━━ 🔀highfr ┃ ┣━━ 📉fm2 ┃ ┗━━ 📉fm1 ┗━━ 🔀mediumfr ┣━━ 📉fm2 ┗━━ 📉fm1
In [17]:
Copied!
sc = case.scenarios_maker()
sc
sc = case.scenarios_maker()
sc
Out[17]:
[{'DC2': ['base'],
'DC3': ['base'],
'DC4': ['base'],
'DC5': ['base'],
'well-1': ['highfr'],
'well-2': ['highfr']},
{'DC2': ['base'],
'DC3': ['base'],
'DC4': ['base'],
'DC5': ['base'],
'well-1': ['mediumfr'],
'well-2': ['highfr']},
{'DC2': ['base'],
'DC3': ['base'],
'DC4': ['base'],
'DC5': ['base'],
'well-1': ['highfr'],
'well-2': ['mediumfr']},
{'DC2': ['base'],
'DC3': ['base'],
'DC4': ['base'],
'DC5': ['base'],
'well-1': ['mediumfr'],
'well-2': ['mediumfr']}]
When the reduced key argument is set, it triggers the Generalized Subset Designed (GSD) (Well explained on PyDOE2 Documentation).
Generate Forecast.¶
Those scenarios are a list of dictionary ready to be passed to generate_forecast method.
In [18]:
Copied!
fwn= case.generate_forecast(wells=sc[3],freq_output='A',iter=50, seed=21)
#fwn
sns.lineplot(data=fwn, x=fwn.index.to_timestamp(), y='oil_rate', hue='well',style='scenario',palette='crest')
fwn= case.generate_forecast(wells=sc[3],freq_output='A',iter=50, seed=21)
#fwn
sns.lineplot(data=fwn, x=fwn.index.to_timestamp(), y='oil_rate', hue='well',style='scenario',palette='crest')
Out[18]:
<AxesSubplot:xlabel='date', ylabel='oil_rate'>
Generate Cashflow¶
Once a forecast is generated the generate_cashflow method can be called
In [19]:
Copied!
#cwn= lp.generate_cashflow(wells={'well-1':['mediumfr'],'well-2':['mediumfr']},freq_output='A')
cwn= case.generate_cashflow(wells=sc[3],freq_output='A')
print(len(cwn))
#cwn= lp.generate_cashflow(wells={'well-1':['mediumfr'],'well-2':['mediumfr']},freq_output='A')
cwn= case.generate_cashflow(wells=sc[3],freq_output='A')
print(len(cwn))
50
Notice that the resulting cashflow for the third scenario is a list of 50 CashflowModel due to the iter key argument was set with this number. What it does is to generate 50 samples of the probabilistic varibales if they exist in the scenarios and broadcast results with the scenarios that does not have any probabilistic variables.
Let's see the first cashflow model
In [20]:
Copied!
fcf_0= cwn[30].fcf()
#Show cashflows in Millons of dollars
print(fcf_0.multiply(1e-6).round(2))
fcf_0= cwn[30].fcf()
#Show cashflows in Millons of dollars
print(fcf_0.multiply(1e-6).round(2))
income_fdp_field-DC2-base-pdp income_fdp_field-DC3-base-pdp \
2021 1.12 3.53
2022 0.02 2.48
2023 0.00 1.99
2024 0.00 1.08
2025 0.00 0.53
2026 0.00 0.00
2027 0.00 0.00
2028 0.00 0.00
2029 0.00 0.00
2030 0.00 0.00
income_fdp_field-DC4-base-pdp income_fdp_field-DC5-base-pdp \
2021 0.95 2.09
2022 0.00 1.44
2023 0.00 0.53
2024 0.00 0.00
2025 0.00 0.00
2026 0.00 0.00
2027 0.00 0.00
2028 0.00 0.00
2029 0.00 0.00
2030 0.00 0.00
income_fdp_field-well-1-mediumfr-fm1 \
2021 0.00
2022 30.59
2023 18.09
2024 11.99
2025 6.87
2026 4.20
2027 0.00
2028 0.00
2029 0.00
2030 0.00
income_fdp_field-well-1-mediumfr-fm2 \
2021 0.00
2022 0.00
2023 0.00
2024 0.00
2025 0.00
2026 20.77
2027 8.38
2028 4.40
2029 3.04
2030 0.00
income_fdp_field-well-2-mediumfr-fm2 \
2021 0.00
2022 20.56
2023 11.29
2024 5.92
2025 0.00
2026 0.00
2027 0.00
2028 0.00
2029 0.00
2030 0.00
income_fdp_field-well-2-mediumfr-fm1 total_income \
2021 0.00 7.69
2022 0.00 55.09
2023 0.00 31.90
2024 35.99 54.98
2025 14.48 21.88
2026 10.07 35.04
2027 6.28 14.66
2028 4.10 8.50
2029 3.22 6.26
2030 0.00 0.01
fix_opex_fdp_field-DC2-base-pdp ... \
2021 -0.13 ...
2022 -0.13 ...
2023 0.00 ...
2024 0.00 ...
2025 0.00 ...
2026 0.00 ...
2027 0.00 ...
2028 0.00 ...
2029 0.00 ...
2030 0.00 ...
capex_drill_fdp_field-well-1-mediumfr-fm1 \
2021 0.0
2022 -4.0
2023 0.0
2024 0.0
2025 0.0
2026 0.0
2027 0.0
2028 0.0
2029 0.0
2030 0.0
capex_wo_fdp_field-well-1-mediumfr-fm2 \
2021 0.0
2022 0.0
2023 0.0
2024 0.0
2025 0.0
2026 -0.6
2027 0.0
2028 0.0
2029 0.0
2030 0.0
capex_abandon_fdp_field-well-1-mediumfr-fm2 \
2021 0.0
2022 0.0
2023 0.0
2024 0.0
2025 0.0
2026 0.0
2027 0.0
2028 0.0
2029 0.0
2030 -0.2
capex_drill_fdp_field-well-2-mediumfr-fm2 \
2021 0.0
2022 -4.0
2023 0.0
2024 0.0
2025 0.0
2026 0.0
2027 0.0
2028 0.0
2029 0.0
2030 0.0
capex_wo_fdp_field-well-2-mediumfr-fm1 \
2021 0.0
2022 0.0
2023 0.0
2024 -0.6
2025 0.0
2026 0.0
2027 0.0
2028 0.0
2029 0.0
2030 0.0
capex_abandon_fdp_field-well-2-mediumfr-fm1 buy total_capex fcf \
2021 0.0 -15.5 -15.7 -9.65
2022 0.0 0.0 -8.2 40.27
2023 0.0 0.0 -0.2 28.19
2024 0.0 0.0 -0.6 48.86
2025 0.0 0.0 -0.2 18.78
2026 0.0 0.0 -0.6 30.34
2027 0.0 0.0 0.0 12.57
2028 0.0 0.0 0.0 6.97
2029 0.0 0.0 0.0 5.02
2030 -0.2 0.0 -0.4 -0.66
cum_fcf
2021 -9.65
2022 30.62
2023 58.81
2024 107.67
2025 126.45
2026 156.79
2027 169.36
2028 176.33
2029 181.35
2030 180.69
[10 rows x 40 columns]
Here there are two additional features in the Cashflow Parameter definition.
- If you want to assign an additional Capex to the project that does not assigned to a single well but the case itself, for example a Purchase price, infraestructure investment, etc..., you can declare a
CashFlowParamwith the key argumentgeneralset to True. What it does is not to pass the cashflow parameter to each well instead make a general cashflow for the model.
In [21]:
Copied!
case.cashflow_params[3]
case.cashflow_params[3]
Out[21]:
CashFlowParams(name='buy', wi=1.0, periods=1, value=-15500000.0, target=<TargetEnum.capex: 'capex'>, multiply=None, agg='mean', depends=False, iter=1, general=True, freq_value=None)
- The income cashflow_param is a Wiener Class, especifically a
MeanReversioninstance. In this case the Oil Price is modeled by the Mean Reversion Model which gives a different time series on each of the 50 iterations.
In [22]:
Copied!
case.cashflow_params[2]
case.cashflow_params[2]
Out[22]:
CashFlowParams(name='income', wi=0.92, periods=None, value=MeanReversion(initial_condition=60.0, ti=datetime.date(2021, 4, 1), steps=11, processes=1, generator=ProbVar(dist='norm', kw={'loc': 0.0, 'scale': 13.13}, factor=1.0, seed=None), freq_input=<FreqEnum.A: 'A'>, freq_output='D', m=46.77, eta=0.112653), target=<TargetEnum.income: 'income'>, multiply='oil_volume', agg='mean', depends=False, iter=1, general=False, freq_value=None)
The next plot shows what could be the oil price iterations for the 50 cases
In [23]:
Copied!
oil_price_model = MeanReversion(
initial_condition = 66,
ti = date(2021,1,1),
generator = {'dist':'norm','kw':{'loc':0,'scale':5.13}},
m=46.77,
eta=0.112652,
freq_input = 'A'
)
oil_price = oil_price_model.generate(12,50, freq_output='A', seed=21)
oil_price.plot(legend=False)
oil_price_model = MeanReversion(
initial_condition = 66,
ti = date(2021,1,1),
generator = {'dist':'norm','kw':{'loc':0,'scale':5.13}},
m=46.77,
eta=0.112652,
freq_input = 'A'
)
oil_price = oil_price_model.generate(12,50, freq_output='A', seed=21)
oil_price.plot(legend=False)
Out[23]:
<AxesSubplot:>
Get the Net present value for the 50 cases. Plot the distribution of the scenarios
In [24]:
Copied!
npv = case.npv([0.15], freq_rate='A', freq_cashflow='A')/1e6
print(npv['npv'].quantile([0.1,0.5,0.9]))
sns.displot(npv['npv'].values, kde=True)
npv = case.npv([0.15], freq_rate='A', freq_cashflow='A')/1e6
print(npv['npv'].quantile([0.1,0.5,0.9]))
sns.displot(npv['npv'].values, kde=True)
0.1 23.561044 0.5 59.131838 0.9 95.542097 Name: npv, dtype: float64
Out[24]:
<seaborn.axisgrid.FacetGrid at 0x7f47ff6ef580>
Plot any of the cases
In [25]:
Copied!
cwn[38].plot(cum=True,format='m')
cwn[38].plot(cum=True,format='m')
/home/scuervo/Documents/dev/apps/dcapy/dcapy/cashflow/cashflow.py:351: UserWarning: FixedFormatter should only be used together with FixedLocator grax.set_yticklabels([fmt.format(i/format_dict[format]['factor']) for i in ticks]) /home/scuervo/Documents/dev/apps/dcapy/dcapy/cashflow/cashflow.py:359: UserWarning: FixedFormatter should only be used together with FixedLocator spax.set_yticklabels([fmt.format(i/format_dict[format]['factor']) for i in ticks_cum])
In [26]:
Copied!
with open('YML_example1.yml','r') as file:
case_dict = yaml.load(file)
lp = WellsGroup(**case_dict)
with open('YML_example1.yml','r') as file:
case_dict = yaml.load(file)
lp = WellsGroup(**case_dict)
<ipython-input-26-ebe3c98787ec>:2: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details. case_dict = yaml.load(file)
In [27]:
Copied!
fwn= lp.generate_forecast(wells=sc[3],freq_output='A',iter=2, seed=21)
cwn= lp.generate_cashflow(wells=sc[3],freq_output='A')
fwn= lp.generate_forecast(wells=sc[3],freq_output='A',iter=2, seed=21)
cwn= lp.generate_cashflow(wells=sc[3],freq_output='A')
In [28]:
Copied!
cwn[0].fcf()['cum_fcf'].values
cwn[0].fcf()['cum_fcf'].values
Out[28]:
array([-9.65263710e+06, 2.09036981e+07, 4.34884735e+07, 7.09743677e+07,
8.67308430e+07, 9.95260309e+07, 1.05163157e+08, 1.08640398e+08,
1.13126400e+08, 1.12473869e+08])
In [29]:
Copied!
npv_list = []
for i,s in enumerate(sc):
fwn = case.generate_forecast(wells=s,freq_output='A',iter=50, seed=21)
cwn = case.generate_cashflow(wells=s,freq_output='A')
npv = case.npv([0.15], freq_rate='A', freq_cashflow='A')/1e6
npv['sc'] = i
npv_list.append(npv)
npv_df = pd.concat(npv_list, axis=0)
print(npv['npv'].quantile([0.1,0.5,0.9]))
sns.displot(npv_df, x='npv', hue='sc', multiple="stack")
npv_list = []
for i,s in enumerate(sc):
fwn = case.generate_forecast(wells=s,freq_output='A',iter=50, seed=21)
cwn = case.generate_cashflow(wells=s,freq_output='A')
npv = case.npv([0.15], freq_rate='A', freq_cashflow='A')/1e6
npv['sc'] = i
npv_list.append(npv)
npv_df = pd.concat(npv_list, axis=0)
print(npv['npv'].quantile([0.1,0.5,0.9]))
sns.displot(npv_df, x='npv', hue='sc', multiple="stack")
0.1 -4.309976 0.5 3.403842 0.9 9.881238 Name: npv, dtype: float64 0.1 10.255532 0.5 30.245220 0.9 52.281651 Name: npv, dtype: float64 0.1 9.219426 0.5 30.774217 0.9 52.175129 Name: npv, dtype: float64 0.1 23.561044 0.5 59.131838 0.9 95.542097 Name: npv, dtype: float64
Out[29]:
<seaborn.axisgrid.FacetGrid at 0x7f4802035d00>
In [ ]:
Copied!