def reset(self, identifier, coderoot):
"""Resets the writer to work with a new executable."""
#A method writer has all the guts needed to generate the executable
#We are just wrapping it and making sure the variable values get
#recorded to temporary files for comparison etc.
self.writer = MethodWriter(identifier, self.parser, self.testgen)
self.identifier = identifier
if self.libraryroot is not None:
self.folders[identifier] = path.abspath(self.libraryroot)
elif self.writer.group is not None and self.writer.group.staging is not None:
#Determine the absolute path using the relative path specified in the group
#staging directory attribute.
from fortpy.tramp import coderelpath
self.folders[identifier] = coderelpath(coderoot,
self.writer.group.staging)
else:
raise ValueError(
"Can't determine the staging directory for running the tests.")
self._needs = None
#The calling methods need to know whether the context (i.e. the staging folder)
#of this executable writer has changed. This variable tracks that.
newstage = self.folder != self.folders[identifier]
self.folder = path.join(self.folders[identifier], identifier)
#Create the directory for the executable files to be copied and written to.
if not path.exists(self.folder):
msg.okay("EXEC DIR: create {}".format(self.folder))
mkdir(self.folder)
return newstage
def reset(self, identifier, coderoot):
"""Resets the writer to work with a new executable."""
#A method writer has all the guts needed to generate the executable
#We are just wrapping it and making sure the variable values get
#recorded to temporary files for comparison etc.
self.writer = MethodWriter(identifier, self.parser, self.testgen)
self.identifier = identifier
if self.libraryroot is not None:
self.folders[identifier] = path.abspath(self.libraryroot)
elif self.writer.group is not None and self.writer.group.staging is not None:
#Determine the absolute path using the relative path specified in the group
#staging directory attribute.
from fortpy.tramp import coderelpath
self.folders[identifier] = coderelpath(coderoot, self.writer.group.staging)
else:
raise ValueError("Can't determine the staging directory for running the tests.")
self._needs = None
#The calling methods need to know whether the context (i.e. the staging folder)
#of this executable writer has changed. This variable tracks that.
newstage = self.folder != self.folders[identifier]
self.folder = path.join(self.folders[identifier], identifier)
#Create the directory for the executable files to be copied and written to.
if not path.exists(self.folder):
msg.okay("EXEC DIR: create {}".format(self.folder))
mkdir(self.folder)
return newstage
class ExecutableGenerator(object):
"""Generates a fortran executable to perform unit tests for a given
subroutine or function.
:arg parser: an instance of the code parser for inter-module access
:arg folder: the folder in which to generate the code files and execs.
:arg testgen: the TestGenerator instance that owns this executable generator.
"""
def __init__(self, parser, folder, testgen):
#We need the code parser because the pre-reqs specified may be in other
#modules and may have pre-reqs of their own. This way, we can find
#them all easily.
self.parser = parser
self.libraryroot = folder
"""Specifies a staging directory to override the one specified in the <group>
tag (if one was specified there)."""
self.testgen = testgen
self.folders = {}
"""Dictionary of full paths to the staging directory for each unit test
being generated by the framework.
"""
self.folder = None
"""The full path to the *current* staging directory for the executable whose
tests are being generated.
"""
self._needs = None
def needs(self):
"""Returns a list of all the modules that this executable needs to
run correctly."""
if self._needs is None:
self._needs = self._calc_needs()
return self._needs
def _calc_needs(self):
"""Calculates a list of all the modules that this executable needs to
run correctly."""
result = []
for modk in self.writer.uses():
if modk not in result:
result.append(modk)
#We also need to look up the dependencies of each of these modules
recursed = list(result)
for i in range(len(result)):
module = result[i]
self._process_module_needs(module, i, recursed)
return recursed
def _process_module_needs(self, module, i, result):
"""Adds the module and its dependencies to the result list."""
#Some code might decide to use the fortpy module methods for general
#development, ignore it since we know it will be present in the end.
if module == "fortpy":
return
#See if the parser has alread loaded this module.
if module not in self.parser.modules:
self.parser.load_dependency(module, True, True, False)
#It is possible that the parser couldn't find it, if so
#we can't create the executable!
if module in self.parser.modules:
modneeds = self.parser.modules[module].needs
for modn in modneeds:
if modn not in result:
#Since this module depends on the other, insert the other
#above it in the list.
result.insert(i, modn)
else:
x = result.index(modn)
if x > i:
#We need to move this module higher up in the food chain
#because it is needed sooner.
result.remove(modn)
result.insert(i, modn)
newi = result.index(modn)
self._process_module_needs(modn, newi, result)
else:
raise ValueError("unable to find module {}.".format(module))
def reset(self, identifier, coderoot):
"""Resets the writer to work with a new executable."""
#A method writer has all the guts needed to generate the executable
#We are just wrapping it and making sure the variable values get
#recorded to temporary files for comparison etc.
self.writer = MethodWriter(identifier, self.parser, self.testgen)
self.identifier = identifier
if self.libraryroot is not None:
self.folders[identifier] = path.abspath(self.libraryroot)
elif self.writer.group is not None and self.writer.group.staging is not None:
#Determine the absolute path using the relative path specified in the group
#staging directory attribute.
from fortpy.tramp import coderelpath
self.folders[identifier] = coderelpath(coderoot,
self.writer.group.staging)
else:
raise ValueError(
"Can't determine the staging directory for running the tests.")
self._needs = None
#The calling methods need to know whether the context (i.e. the staging folder)
#of this executable writer has changed. This variable tracks that.
newstage = self.folder != self.folders[identifier]
self.folder = path.join(self.folders[identifier], identifier)
#Create the directory for the executable files to be copied and written to.
if not path.exists(self.folder):
msg.okay("EXEC DIR: create {}".format(self.folder))
mkdir(self.folder)
return newstage
def write(self, testid, coderoot):
"""Writes the fortran program file for the executable specified.
:arg testid: the identifier of the test to construct the executable for.
"""
lines = []
identifier = self.writer.tests[testid].identifier
#First off, we need to start the program and set the module dependencies.
lines.append("!!<summary>Auto-generated unit test for {}\n".format(
self.identifier))
lines.append("!!using FORTPY. Generated on {}.\n".format(
datetime.now()))
lines.append("!!{}</summary>\n".format(
self.writer.tests[testid].description))
lines.append("PROGRAM UNITTEST_{}\n".format(
self.writer.finders[testid].executable.name))
lines.append(self._get_uses(testid))
lines.append(" implicit none\n")
#Next add the variable declarations and initializations and the calls
#to execute the pre-req methods and the one we are trying to test.
lines.append(self.writer.lines(testid, coderoot))
lines.append("\nEND PROGRAM UNITTEST_{}".format(
self.writer.finders[testid].executable.name))
with open(path.join(self.folder, "{}.f90".format(identifier)),
'w') as f:
f.writelines(lines)
def makefile(self, identifier):
"""Generates a makefile to create the unit testing executable
for the specified test identifier.
:arg identifier: the id of the test that this executable should be made for.
"""
allneeds = self.needs()
#We need to see whether to include the pre-compiler directive or not.
precompile = False
for needed in allneeds:
if self.parser.modules[needed].precompile:
precompile = True
break
lines = []
makepath = path.join(self.folder, "Makefile.{}".format(identifier))
makefile(identifier,
allneeds,
makepath,
self.identifier,
precompile,
parser=self.parser)
def _get_uses(self, testid):
"""Gets a list of use statements to add to the program code."""
#The writer can extract a dictionary of module dependencies for us.
alluses = self.writer.uses()
#Now we just need to generate the code statements.
uselist = []
for module in alluses:
if module == self.writer.finders[testid].module.name:
uselist.append("use {}".format(module))
else:
uselist.append("use {}, only: {}".format(
module, ", ".join(alluses[module])))
#Last of all, we need to append the module that handles interaction with
#the fortran results and the python testing framework.
if "fortpy" not in alluses:
uselist.append("use fortpy\n")
return self._tabjoin(uselist)
def _tabjoin(self, values):
"""Joins a list of values with \t\n."""
return " {}".format("\n ".join(values))
class ExecutableGenerator(object):
"""Generates a fortran executable to perform unit tests for a given
subroutine or function.
:arg parser: an instance of the code parser for inter-module access
:arg folder: the folder in which to generate the code files and execs.
:arg testgen: the TestGenerator instance that owns this executable generator.
"""
def __init__(self, parser, folder, testgen):
#We need the code parser because the pre-reqs specified may be in other
#modules and may have pre-reqs of their own. This way, we can find
#them all easily.
self.parser = parser
self.libraryroot = folder
"""Specifies a staging directory to override the one specified in the <group>
tag (if one was specified there)."""
self.testgen = testgen
self.folders = {}
"""Dictionary of full paths to the staging directory for each unit test
being generated by the framework.
"""
self.folder = None
"""The full path to the *current* staging directory for the executable whose
tests are being generated.
"""
self._needs = None
def needs(self):
"""Returns a list of all the modules that this executable needs to
run correctly."""
if self._needs is None:
self._needs = self._calc_needs()
return self._needs
def _calc_needs(self):
"""Calculates a list of all the modules that this executable needs to
run correctly."""
result = []
for modk in self.writer.uses():
if modk not in result:
result.append(modk)
#We also need to look up the dependencies of each of these modules
recursed = list(result)
for i in range(len(result)):
module = result[i]
self._process_module_needs(module, i, recursed)
return recursed
def _process_module_needs(self, module, i, result):
"""Adds the module and its dependencies to the result list."""
#Some code might decide to use the fortpy module methods for general
#development, ignore it since we know it will be present in the end.
if module == "fortpy":
return
#See if the parser has alread loaded this module.
if module not in self.parser.modules:
self.parser.load_dependency(module, True, True, False)
#It is possible that the parser couldn't find it, if so
#we can't create the executable!
if module in self.parser.modules:
modneeds = self.parser.modules[module].needs
for modn in modneeds:
if modn not in result:
#Since this module depends on the other, insert the other
#above it in the list.
result.insert(i, modn)
else:
x = result.index(modn)
if x > i:
#We need to move this module higher up in the food chain
#because it is needed sooner.
result.remove(modn)
result.insert(i, modn)
newi = result.index(modn)
self._process_module_needs(modn, newi, result)
else:
raise ValueError("unable to find module {}.".format(module))
def reset(self, identifier, coderoot):
"""Resets the writer to work with a new executable."""
#A method writer has all the guts needed to generate the executable
#We are just wrapping it and making sure the variable values get
#recorded to temporary files for comparison etc.
self.writer = MethodWriter(identifier, self.parser, self.testgen)
self.identifier = identifier
if self.libraryroot is not None:
self.folders[identifier] = path.abspath(self.libraryroot)
elif self.writer.group is not None and self.writer.group.staging is not None:
#Determine the absolute path using the relative path specified in the group
#staging directory attribute.
from fortpy.tramp import coderelpath
self.folders[identifier] = coderelpath(coderoot, self.writer.group.staging)
else:
raise ValueError("Can't determine the staging directory for running the tests.")
self._needs = None
#The calling methods need to know whether the context (i.e. the staging folder)
#of this executable writer has changed. This variable tracks that.
newstage = self.folder != self.folders[identifier]
self.folder = path.join(self.folders[identifier], identifier)
#Create the directory for the executable files to be copied and written to.
if not path.exists(self.folder):
msg.okay("EXEC DIR: create {}".format(self.folder))
mkdir(self.folder)
return newstage
def write(self, testid, coderoot):
"""Writes the fortran program file for the executable specified.
:arg testid: the identifier of the test to construct the executable for.
"""
lines = []
identifier = self.writer.tests[testid].identifier
#First off, we need to start the program and set the module dependencies.
lines.append("!!<summary>Auto-generated unit test for {}\n".format(
self.identifier))
lines.append("!!using FORTPY. Generated on {}.\n".format(datetime.now()))
lines.append("!!{}</summary>\n".format(self.writer.tests[testid].description))
lines.append("PROGRAM UNITTEST_{}\n".format(self.writer.finders[testid].executable.name))
lines.append(self._get_uses(testid))
lines.append(" implicit none\n")
#Next add the variable declarations and initializations and the calls
#to execute the pre-req methods and the one we are trying to test.
lines.append(self.writer.lines(testid, coderoot))
lines.append("\nEND PROGRAM UNITTEST_{}".format(self.writer.finders[testid].executable.name))
with open(path.join(self.folder, "{}.f90".format(identifier)), 'w') as f:
f.writelines(lines)
def makefile(self, identifier):
"""Generates a makefile to create the unit testing executable
for the specified test identifier.
:arg identifier: the id of the test that this executable should be made for.
"""
allneeds = self.needs()
#We need to see whether to include the pre-compiler directive or not.
precompile = False
for needed in allneeds:
if self.parser.modules[needed].precompile:
precompile = True
break
lines = []
makepath = path.join(self.folder, "Makefile.{}".format(identifier))
makefile(identifier, allneeds, makepath, self.identifier, precompile, parser=self.parser)
def _get_uses(self, testid):
"""Gets a list of use statements to add to the program code."""
#The writer can extract a dictionary of module dependencies for us.
alluses = self.writer.uses()
#Now we just need to generate the code statements.
uselist = []
for module in alluses:
if module == self.writer.finders[testid].module.name:
uselist.append("use {}".format(module))
else:
uselist.append("use {}, only: {}".format(module, ", ".join(alluses[module])))
#Last of all, we need to append the module that handles interaction with
#the fortran results and the python testing framework.
if "fortpy" not in alluses:
uselist.append("use fortpy\n")
return self._tabjoin(uselist)
def _tabjoin(self, values):
"""Joins a list of values with \t\n."""
return " {}".format("\n ".join(values))
class ExecutableGenerator(object):
"""Generates a fortran executable to perform unit tests for a given
subroutine or function.
:arg parser: an instance of the code parser for inter-module access
:arg folder: the folder in which to generate the code files and execs.
:arg testgen: the TestGenerator instance that owns this executable generator.
"""
def __init__(self, parser, folder, testgen):
#We need the code parser because the pre-reqs specified may be in other
#modules and may have pre-reqs of their own. This way, we can find
#them all easily.
self.parser = parser
self.libraryroot = folder
"""Specifies a staging directory to override the one specified in the <group>
tag (if one was specified there)."""
self.testgen = testgen
self.folders = {}
"""Dictionary of full paths to the staging directory for each unit test
being generated by the framework.
"""
self.folder = None
"""The full path to the *current* staging directory for the executable whose
tests are being generated.
"""
self._needs = None
def needs(self):
"""Returns a list of all the modules that this executable needs to
run correctly."""
if self._needs is None:
from fortpy.code import order_module_dependencies
self._needs = order_module_dependencies(self.writer.uses(), self.parser)
return self._needs
def reset(self, identifier, coderoot):
"""Resets the writer to work with a new executable."""
#A method writer has all the guts needed to generate the executable
#We are just wrapping it and making sure the variable values get
#recorded to temporary files for comparison etc.
self.writer = MethodWriter(identifier, self.parser, self.testgen,
stagedir=self.libraryroot)
self.identifier = identifier
if self.libraryroot is not None:
self.folders[identifier] = path.abspath(self.libraryroot)
elif self.writer.group is not None and self.writer.group.staging is not None:
#Determine the absolute path using the relative path specified in the group
#staging directory attribute.
from fortpy.tramp import coderelpath
self.folders[identifier] = coderelpath(coderoot, self.writer.group.staging)
else:
raise ValueError("Can't determine the staging directory for running the tests.")
self._needs = None
#The calling methods need to know whether the context (i.e. the staging folder)
#of this executable writer has changed. This variable tracks that.
newstage = self.folder != self.folders[identifier]
self.folder = path.join(self.folders[identifier], identifier)
#Create the directory for the executable files to be copied and written to.
if not path.exists(self.folder):
msg.okay("EXEC DIR: create {}".format(self.folder))
mkdir(self.folder)
return newstage
def write(self, testid, coderoot):
"""Writes the fortran program file for the executable specified.
:arg testid: the identifier of the test to construct the executable for.
"""
lines = []
identifier = self.writer.tests[testid].identifier
#First off, we need to start the program and set the module dependencies.
lines.append("!!<summary>Auto-generated unit test for {}\n".format(
self.identifier))
lines.append("!!using FORTPY. Generated on {}.\n".format(datetime.now()))
lines.append("!!{}</summary>\n".format(self.writer.tests[testid].description))
lines.append("PROGRAM UNITTEST_{}\n".format(self.writer.finders[testid].executable.name))
lines.append(self._get_uses(testid))
lines.append(" implicit none\n")
#Next add the variable declarations and initializations and the calls
#to execute the pre-req methods and the one we are trying to test.
lines.append(self.writer.lines(testid, coderoot))
lines.append("\nEND PROGRAM UNITTEST_{}".format(self.writer.finders[testid].executable.name))
with open(path.join(self.folder, "{}.f90".format(identifier)), 'w') as f:
f.writelines(lines)
def makefile(self, identifier):
"""Generates a makefile to create the unit testing executable
for the specified test identifier.
:arg identifier: the id of the test that this executable should be made for.
"""
allneeds = self.needs()
#We need to see whether to include the pre-compiler directive or not.
precompile = False
for needed in allneeds:
if needed=="fortpy" or needed=="fpy_auxiliary":
continue
if self.parser.modules[needed].precompile:
precompile = True
break
lines = []
#Determine whether the compilation should produce headers on the output screen.
from fortpy.msg import will_print
verbose = will_print(2)
makepath = path.join(self.folder, "Makefile.{}".format(identifier))
makefile(identifier, allneeds, makepath, self.identifier, precompile,
parser=self.parser, inclfpyaux=self.writer.autoclass, verbose=verbose)
def _get_uses(self, testid):
"""Gets a list of use statements to add to the program code."""
#The writer can extract a dictionary of module dependencies for us.
alluses = self.writer.uses()
#Now we just need to generate the code statements.
uselist = []
for module in alluses:
if module == self.writer.finders[testid].module.name:
uselist.append("use {}".format(module))
elif module == "fpy_auxiliary":
uselist.append("use fpy_auxiliary")
else:
uselist.append("use {}, only: {}".format(module, ", ".join(alluses[module])))
#Last of all, we need to append the module that handles interaction with
#the fortran results and the python testing framework.
if "fortpy" not in alluses:
uselist.append("use fortpy\n")
return self._tabjoin(uselist)
def _tabjoin(self, values):
"""Joins a list of values with \t\n."""
return " {}".format("\n ".join(values))
