Skip to content
Snippets Groups Projects
Commit 0ad189aa authored by David Hendriks's avatar David Hendriks
Browse files

fixing the api log issue when running the tests

parent 09549a7e
No related branches found
No related tags found
No related merge requests found
%% Cell type:markdown id:a544d28c-c2e1-4c6a-b55b-8caec440743f tags:
# Running individual systems with binarycpython
This notebook will show you how to run single systems and analyze their results.
It can be useful to have some functions to quickly run a single system to e.g. inspect what evolutionary steps a specific system goes through, to plot the mass loss evolution of a single star, etc.
%% Cell type:markdown id:dd5d9ec7-5791-45f1-afbd-225947e2a583 tags:
## Single system with run_wrapper
The simplest method to run a single system is to use the run_system wrapper. This function deals with setting up the argument string, makes sure all the required parameters are included and handles setting and cleaning up the custom logging functionality (see notebook_custom_logging).
As arguments to this function we can add any of the parameters that binary_c itself actually knows, as well as:
- custom_logging_code: string containing a print statement that binary_c can use to print information
- log_filename: path of the logfile that binary_c generates
- parse_function: function that handles parsing the output of binary-c
%% Cell type:code id:e32dcdee tags:
``` python
from binarycpython.utils.functions import temp_dir
TMP_DIR = temp_dir("notebooks", "notebook_individual_systems")
```
%% Cell type:code id:425efed3-d8e3-432d-829e-41d8ebe05162 tags:
``` python
from binarycpython.utils.run_system_wrapper import run_system
# help(run_system) # Uncomment to see the docstring
```
%% Cell type:code id:b2abab48-433d-4936-8434-14804c52c9f6 tags:
``` python
output = run_system(M_1=1)
print(output)
```
%% Output
SINGLE_STAR_LIFETIME 1 12462
SINGLE_STAR_LIFETIME 1 12461.9
%% Cell type:markdown id:f127a5e4-dc01-4472-9130-8a943c92e8a7 tags:
Lets try adding a log filename now:
%% Cell type:code id:029fc3f2-f09a-49af-a32b-248505738f2e tags:
``` python
output = run_system(M_1=1, log_filename='/tmp/test_logfile.txt')
with open('/tmp/test_logfile.txt', 'r') as f:
log_filename = os.path.join(TMP_DIR, 'test_logfile.txt')
output = run_system(M_1=1, log_filename=log_filename, api_log_filename_prefix=TMP_DIR)
with open(log_filename, 'r') as f:
print(f.read())
```
%% Output
TIME M1 M2 K1 K2 SEP ECC R1/ROL1 R2/ROL2 TYPE RANDOM_SEED=67365 RANDOM_COUNT=0
0.0000 1.000 0.000 1 15 -1 -1.00 0.000 0.000 "INITIAL "
11003.1302 1.000 0.000 2 15 -1 -1.00 0.000 0.000 "OFF_MS"
11003.1302 1.000 0.000 2 15 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
11582.2424 1.000 0.000 3 15 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
12325.1085 0.817 0.000 4 15 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
12457.1300 0.783 0.000 5 15 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
12460.8955 0.774 0.000 6 15 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
12460.8955 0.774 0.000 6 15 -1 -1.00 0.000 0.000 "shrinkAGB"
12461.9514 0.523 0.000 11 15 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
15000.0000 0.523 0.000 11 15 -1 -1.00 0.000 0.000 "MAX_TIME"
TIME M1 M2 K1 K2 SEP PER ECC R1/ROL1 R2/ROL2 TYPE RANDOM_SEED=84984 RANDOM_COUNT=0
0.0000 1.000 0.000 1 15 -1 -1 -1.00 0.000 0.000 "INITIAL "
11003.1302 1.000 0.000 2 15 -1 -1 -1.00 0.000 0.000 "OFF_MS"
11003.1302 1.000 0.000 2 15 -1 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
11582.2424 1.000 0.000 3 15 -1 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
12325.1085 0.817 0.000 4 15 -1 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
12457.1301 0.783 0.000 5 15 -1 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
12460.9983 0.716 0.000 6 15 -1 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
12460.9983 0.716 0.000 6 15 -1 -1 -1.00 0.000 0.000 "shrinkAGB"
12461.9458 0.518 0.000 11 15 -1 -1 -1.00 0.000 0.000 "TYPE_CHNGE"
15000.0000 0.518 0.000 11 15 -1 -1 -1.00 0.000 0.000 "MAX_TIME"
%% Cell type:markdown id:606670f2-3e0a-43c7-a885-006b92fac9d2 tags:
To get more useful output we can include a custom_logging snippet (see notebook_custom_logging):
%% Cell type:code id:e6a23b55-ca42-440d-83ac-e76a24a83a67 tags:
``` python
from binarycpython.utils.custom_logging_functions import binary_c_log_code
# Create the print statement
custom_logging_print_statement = """
Printf("EXAMPLE_MASSLOSS %30.12e %g %g %d\\n",
//
stardata->model.time, // 1
stardata->star[0].mass, //2
stardata->common.zero_age.mass[0], //4
stardata->star[0].stellar_type //5
);
"""
# Generate entire shared lib code around logging lines
custom_logging_code = binary_c_log_code(custom_logging_print_statement)
output = run_system(M_1=1, custom_logging_code=custom_logging_code)
output = run_system(M_1=1, custom_logging_code=custom_logging_code, api_log_filename_prefix=TMP_DIR)
print(output.splitlines()[:4])
```
%% Output
['EXAMPLE_MASSLOSS 0.000000000000e+00 1 1 1', 'EXAMPLE_MASSLOSS 0.000000000000e+00 1 1 1', 'EXAMPLE_MASSLOSS 1.000000000000e-06 1 1 1', 'EXAMPLE_MASSLOSS 2.000000000000e-06 1 1 1']
%% Cell type:markdown id:4c885143-db79-4fed-b4c4-0bd846e24f7d tags:
Now we have some actual output, it is time to create a parse_function which parses the output. Adding a parse_function to the run_system will make run_system run the output of binary_c through the parse_function.
%% Cell type:code id:3822721f-217a-495b-962e-d57137b9e290 tags:
``` python
def parse_function(output):
"""
Example function to parse the output of binary_c
"""
#
column_names = ['time', 'mass', 'initial_mass', 'stellar_type']
value_lines = [column_names]
# Loop over output
for line in output.splitlines():
# Select the lines starting with the header we chose
if line.startswith("EXAMPLE_MASSLOSS"):
# Split the output and fetch the data
split_line = line.split()
values = [float(el) for el in split_line[1:]]
value_lines.append(values)
return value_lines
# Catch output
output = run_system(M_1=1, custom_logging_code=custom_logging_code, parse_function=parse_function)
print(output[:3])
```
%% Output
[['time', 'mass', 'initial_mass', 'stellar_type'], [0.0, 1.0, 1.0, 1.0], [0.0, 1.0, 1.0, 1.0]]
%% Cell type:markdown id:a551f07f-2eff-4425-9375-267579a581b3 tags:
This output can now be turned into e.g. an Numpy array or Pandas dataframe (my favorite: makes querying the data very easy)
%% Cell type:code id:654a07ed-2a88-46ff-9da0-b7759580f9f3 tags:
``` python
import pandas as pd
# Load data into dataframe
example_df = pd.DataFrame(output)
# Fix column headers
example_df.columns = example_df.iloc[0]
example_df = example_df.drop(example_df.index[0])
print(example_df)
```
%% Output
0 time mass initial_mass stellar_type
1 0 1 1 1
2 0 1 1 1
3 1e-06 1 1 1
4 2e-06 1 1 1
5 3e-06 1 1 1
... ... ... ... ...
1612 12461.8 0.577754 1 6
1613 12462 0.522806 1 11
1614 13462 0.522806 1 11
1615 14462 0.522806 1 11
1616 15000 0.522806 1 11
1617 12461.8 0.546683 1 6
1618 12461.9 0.517749 1 11
1619 13461.9 0.517749 1 11
1620 14461.9 0.517749 1 11
1621 15000 0.517749 1 11
[1616 rows x 4 columns]
[1621 rows x 4 columns]
%% Cell type:markdown id:325c2ce6-f9a1-46b7-937f-84040e1252cf tags:
## Single system via population object
When setting up your population object (see notebook_population), and configuring all the parameters, it is possible to run a single system using that same configuration. It will use the parse_function if set, and running a single system is a good method to test if everything works accordingly.
%% Cell type:code id:4a98ffca-1b72-4bb8-8df1-3bf3187d882f tags:
``` python
from binarycpython.utils.grid import Population
# help(Population) # Uncomment to see the docstring
```
%% Cell type:markdown id:7e2c2ef0-3db2-46a6-8c85-9b6cf720eb6a tags:
First, let's try this without any custom logging or parsing functionality
%% Cell type:code id:bff1cc2e-6b32-4ba0-879f-879ffbabd223 tags:
``` python
# Create the population object
example_pop = Population()
# Set some parameters
example_pop.set(
verbosity=1
verbosity=1,
api_log_filename_prefix=TMP_DIR
)
example_pop.set(
M_1=10
)
# get output and print
output = example_pop.evolve_single()
print(output)
```
%% Output
adding: M_1=10 to BSE_options
Creating and loading custom logging functionality
Running binary_c M_1 10
Cleaning up the custom logging stuff. type: single
SINGLE_STAR_LIFETIME 10 27.7358
SINGLE_STAR_LIFETIME 10 28.4838
%% Cell type:markdown id:ae01fa35-f8b1-4a40-bfb2-b9e872cae0e7 tags:
Now lets add some actual output with the custom logging
%% Cell type:code id:dd748bab-b57e-4129-8350-9ea11fa179d0 tags:
``` python
custom_logging_print_statement = """
Printf("EXAMPLE_MASSLOSS %30.12e %g %g %g %d\\n",
//
stardata->model.time, // 1
stardata->star[0].mass, //2
stardata->previous_stardata->star[0].mass, //3
stardata->common.zero_age.mass[0], //4
stardata->star[0].stellar_type //5
);
"""
example_pop.set(C_logging_code=custom_logging_print_statement)
# get output and print
output = example_pop.evolve_single()
print(output.splitlines()[:4])
print('\n'.join(output.splitlines()[:4]))
```
%% Output
adding: C_logging_code=
Printf("EXAMPLE_MASSLOSS %30.12e %g %g %g %d\n",
//
stardata->model.time, // 1
stardata->star[0].mass, //2
stardata->previous_stardata->star[0].mass, //3
stardata->common.zero_age.mass[0], //4
stardata->star[0].stellar_type //5
);
to grid_options
Creating and loading custom logging functionality
Running binary_c M_1 10
Cleaning up the custom logging stuff. type: single
Removed /tmp/binary_c_python/custom_logging/libcustom_logging_eac2dfc438a14e5a9f5be98b1b6b4294.so
['EXAMPLE_MASSLOSS 0.000000000000e+00 10 0 10 1', 'EXAMPLE_MASSLOSS 0.000000000000e+00 10 10 10 1', 'EXAMPLE_MASSLOSS 1.000000000000e-06 10 10 10 1', 'EXAMPLE_MASSLOSS 2.000000000000e-06 10 10 10 1']
Removed /tmp/binary_c_python/custom_logging/libcustom_logging_273b1d161cc245acb09b7c20ee60be99.so
EXAMPLE_MASSLOSS 0.000000000000e+00 10 0 10 1
EXAMPLE_MASSLOSS 0.000000000000e+00 10 10 10 1
EXAMPLE_MASSLOSS 1.000000000000e-06 10 10 10 1
EXAMPLE_MASSLOSS 2.000000000000e-06 10 10 10 1
%% Cell type:markdown id:588a7d9e-9d64-4b3b-8907-656b905286e8 tags:
Lastly we can add a parse_function to handle parsing the output again.
Because the parse_function will now be part of the population object, it can access information of the object. We need to make a new parse function that is fit for an object: we the arguments now need to be (self, output). Returning the data is useful when running evolve_single(), but won't be used in a population evolution.
%% Cell type:code id:fec39154-cce6-438c-8c2c-509d76b00f34 tags:
``` python
import os
import json
import numpy as np
def object_parse_function(self, output):
"""
Example parse function that can be added to the population object
"""
# We can access object instance information now.
# In this way we can store the results in a file for example.
output_file = os.path.join(self.custom_options['output_dir'], 'example_output.json')
#
column_names = ['time', 'mass', 'initial_mass', 'stellar_type']
value_lines = [column_names]
# Loop over output
for line in output.splitlines():
# Select the lines starting with the header we chose
if line.startswith("EXAMPLE_MASSLOSS"):
# Split the output and fetch the data
split_line = line.split()
values = [float(el) for el in split_line[1:]]
value_lines.append(values)
# Turn into an array
values_array = np.array(value_lines[1:])
# make dict and fill
output_dict = {}
for i in range(len(column_names)):
output_dict[column_names[i]] = list(values_array[:,i])
# Write to file
with open(output_file, 'w') as f:
f.write(json.dumps(output_dict, indent=4))
# Return something anyway
return value_lines
```
%% Cell type:code id:57347512-fd4a-434b-b13c-5e6dbd3ac415 tags:
``` python
example_pop.set(
parse_function=object_parse_function,
output_dir='/tmp/'
output_dir=TMP_DIR,
api_log_filename_prefix=TMP_DIR
)
output = example_pop.evolve_single()
print(output[:4])
# Example of loading the data that was written
with open(os.path.join(example_pop.custom_options['output_dir'], 'example_output.json')) as f:
written_data = json.loads(f.read())
print(written_data.keys())
```
%% Output
adding: parse_function=<function object_parse_function at 0x7f9265091598> to grid_options
<<<< Warning: Key does not match previously known parameter: adding: output_dir=/tmp/ to custom_options >>>>
adding: parse_function=<function object_parse_function at 0x7fbe0c5d1950> to grid_options
<<<< Warning: Key does not match previously known parameter: adding: output_dir=/tmp/binary_c_python/notebooks/notebook_individual_systems to custom_options >>>>
adding: api_log_filename_prefix=/tmp/binary_c_python/notebooks/notebook_individual_systems to BSE_options
Creating and loading custom logging functionality
Running binary_c M_1 10
Running binary_c M_1 10 api_log_filename_prefix /tmp/binary_c_python/notebooks/notebook_individual_systems
Cleaning up the custom logging stuff. type: single
Removed /tmp/binary_c_python/custom_logging/libcustom_logging_e9c2bec7f15541eb847fc6013e48e7ed.so
Removed /tmp/binary_c_python/custom_logging/libcustom_logging_6136984d369d4ba3b18ac029add5ee33.so
[['time', 'mass', 'initial_mass', 'stellar_type'], [0.0, 10.0, 0.0, 10.0, 1.0], [0.0, 10.0, 10.0, 10.0, 1.0], [1e-06, 10.0, 10.0, 10.0, 1.0]]
dict_keys(['time', 'mass', 'initial_mass', 'stellar_type'])
%% Cell type:markdown id:ddc06da3-fc68-4c6f-8067-14ea862b964d tags:
## Single system via API functionality
It is possible to construct your own functionality to run a single system by directly calling the API function to run a system. Under the hood all the other functions and wrappers actually use this API.
There are less failsafes for this method, so this make sure the input is correct and binary_c knows all the arguments you pass in.
for more details on this API function see `notebook_api_functions`
%% Cell type:markdown id:56886792-d379-4eac-b0d4-54508edb39c7 tags:
First we must construct the argument string that we pass to binary_c
%% Cell type:code id:ec48125c-6bf5-48f4-9357-8261800b5d8b tags:
``` python
# For a binary system we need to pass in these arguments
M_1 = 15.0 # Msun
M_2 = 14.0 # Msun
separation = 0 # 0 = ignored, use period
orbital_period = 4530.0 # days
eccentricity = 0.0
metallicity = 0.02
max_evolution_time = 15000 # Myr. You need to include this argument.
api_log_filename_prefix = TMP_DIR
# Here we set up the argument string that is passed to the bindings
argstring = """
binary_c M_1 {M_1} M_2 {M_2} separation {separation} orbital_period {orbital_period} eccentricity {eccentricity} metallicity {metallicity} max_evolution_time {max_evolution_time}
binary_c M_1 {M_1} M_2 {M_2} separation {separation} orbital_period {orbital_period} eccentricity {eccentricity} metallicity {metallicity} max_evolution_time {max_evolution_time} api_log_filename_prefix {api_log_filename_prefix}
""".format(
M_1=M_1,
M_2=M_2,
separation=separation,
orbital_period=orbital_period,
eccentricity=eccentricity,
metallicity=metallicity,
max_evolution_time=max_evolution_time,
api_log_filename_prefix=TMP_DIR
).strip()
from binarycpython import _binary_c_bindings
output = _binary_c_bindings.run_system(argstring)
print(output)
```
%% Output
SINGLE_STAR_LIFETIME 15 14.2383
SINGLE_STAR_LIFETIME 15 14.9947
%% Cell type:markdown id:55c8ea24-0fc0-452c-8121-1e7667433479 tags:
As we can see above, the output is rather empty. But if SINGLE_STAR_LIFETIME is printed we know we caught the output correctly. To get actual output we should have timesteps printed in the `log_every_timestep.c` in binary_c, or add some custom_logging (see notebook_custom_logging).
......
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment