Merge pull request #5 from adelq/JANAF

JANAF

.gitignore

@@ -57,3 +57,8 @@ docs/_build/
 # PyBuilder
 target/
 
+# PyCharm
+.idea/
+
+# JANAF cache directory
+thermochem/JANAF_Cache/

MANIFEST.in

@@ -6,3 +6,4 @@ include TODO
 include ez_setup.py
 
 recursive-include thermochem *.xml
+recursive-include thermochem *.txt

README.rst

@@ -12,6 +12,9 @@ and Thermochemistry. The following tools are provided:
 
 -  Alexander Burcat's database and routines to compute the properties
    of more than 1300 substances from 200K to 6000K.
+
+-  JANAF thermodynamic data pulled from NIST website to compute
+   properties of more than 1200 substances from 0K to 2500K.
 
 -  Simple reactor for modeling combustion of single organic compounds.
 
@@ -28,6 +31,8 @@ This packages needs the following extra modules to run.
 
 -  Scipy >= 0.6.0
 
+-  Pandas >= 0.18.0
+
 -  setuptools >= 0.6
 
 Development

setup.py

@@ -10,7 +10,7 @@
       license="BSD",
       packages=find_packages(),
       include_package_data=True,
-      install_requires=['scipy>=0.6.0', 'numpy>=1.2.1'],
+      install_requires=['scipy>=0.6.0', 'numpy>=1.2.1', 'pandas>=0.17.0'],
       zip_safe=False,
       keywords='thermo chemistry physics',
       classifiers=[

tests/test_janaf.py

@@ -0,0 +1,15 @@
+from thermochem.janaf import Janafdb
+from nose.tools import eq_
+from numpy.testing import assert_allclose
+
+
+def test_titanium():
+    db = Janafdb()
+    ti = db.search('Ti')
+    eq_(len(ti), 88)
+
+
+def test_rutile():
+    db = Janafdb()
+    rutile = db.getphasedata(formula='O2Ti', name='Rutile', phase='cr')
+    assert_allclose(rutile.cp([500, 550, 1800]), [67.203, 68.567, 78.283])

thermochem/__init__.py

@@ -1 +1 @@
-__version__ = '0.5.1'
+__version__ = '0.6.0'

thermochem/janaf.py

@@ -0,0 +1,244 @@
+"""
+This module gets thermodynamic datat from the JANAF database.
+Files are downloaded from the NIST servers as needed and then cached locally.
+
+Zack Gainsforth
+
+Funding by NASA
+"""
+
+from __future__ import division
+from __future__ import print_function
+
+import numpy as np
+import pandas as pd
+from scipy.interpolate import interp1d
+import os
+
+try:
+    # Python 3
+    import urllib.request as urllib2
+except ImportError:
+    # Python 2
+    import urllib2
+
+try:
+    # Python 2
+    from StringIO import StringIO
+except ImportError:
+    # Python 3
+    from io import StringIO
+
+# Universal gas constant R
+R = 8.314472
+
+
+class JanafPhase(object):
+    """
+    Class which is created by Janafdb for a specific phase.
+
+    It reads in the JANAF data file and produces functions which interpolate
+    the thermodynamic constants.
+
+    >>> db = Janafdb()
+    >>> p = db.getphasedata(formula='O2Ti', name='Rutile', phase='cr')
+    >>> print(p.cp([500, 550, 1800]))
+    [ 67.203  68.567  78.283]
+    >>> print(p.S([500, 550, 1800]))
+    [  82.201    88.4565  176.876 ]
+    >>> print(p.gef([500, 550, 1800]))
+    [  57.077   59.704  115.753]
+    >>> print(p.hef([500, 550, 1800]))
+    [  12.562    15.9955  110.022 ]
+    >>> print(p.DeltaH([500, 550, 1800]))
+    [-943670.  -943229.5 -936679. ]
+    >>> print(p.DeltaG([500, 550, 1800]))
+    [-852157.  -843046.5 -621013. ]
+    >>> print(p.logKf([500, 550, 1800]))
+    [ 89.024   80.8125  18.021 ]
+    >>> print(p.cp(50000))
+    Traceback (most recent call last):
+        ...
+    ValueError: A value in x_new is above the interpolation range.
+    """
+
+    def __init__(self, rawdata_text):
+        # Store the raw data text file from NIST.
+        self.rawdata_text = rawdata_text
+
+        # Read the text file into a DataFrame.
+        data = pd.read_csv(
+            StringIO(self.rawdata_text),
+            skiprows=2,
+            header=None,
+            delimiter='[\t\s]*',
+            engine='python')
+        data.columns = ['T', 'Cp', 'S', '[G-H(Tr)]/T', 'H-H(Tr)', 'Delta_fH',
+                        'Delta_fG', 'log(Kf)']
+        self.rawdata = data
+
+        # Sometimes the JANAF files have funky stuff written in them. (Old
+        # school text format...)
+        # Clean it up.
+        for c in data.columns:
+            # We only need to polish up columns that aren't floating point
+            # numbers.
+            if np.issubdtype(data.dtypes[c], np.floating):
+                continue
+            # Change INFINITE to inf
+            data.loc[data[c] == 'INFINITE', c]
+            # Anything else becomes a nan.
+            # Convert to floats.
+            data[c] = pd.to_numeric(data[c], errors='coerce')
+
+        # Convert all units to Joules.
+        data['Delta_fH'] *= 1000
+        data['Delta_fG'] *= 1000
+
+        # Now make interpolatable functions for each of these.
+        self.cp = interp1d(self.rawdata['T'], self.rawdata['Cp'])
+        self.S = interp1d(self.rawdata['T'], self.rawdata['S'])
+        self.gef = interp1d(self.rawdata['T'], self.rawdata['[G-H(Tr)]/T'])
+        self.hef = interp1d(self.rawdata['T'], self.rawdata['H-H(Tr)'])
+        self.DeltaH = interp1d(self.rawdata['T'], self.rawdata['Delta_fH'])
+        self.DeltaG = interp1d(self.rawdata['T'], self.rawdata['Delta_fG'])
+        self.logKf = interp1d(self.rawdata['T'], self.rawdata['log(Kf)'])
+
+        # TODO Deal well with crystal<->liquid transitions which have a below
+        # and above value for Cp, S, etc.
+
+
+class Janafdb(object):
+    """
+    Class that reads the NIST JANAF tables for thermodynamic data.
+
+    Data is initially read from the web servers, and then cached.
+    """
+    VALIDPHASETYPES = ['cr', 'l', 'cr,l', 'g', 'ref', 'cd', 'fl', 'am', 'vit',
+                       'mon', 'pol', 'sln', 'aq', 'sat']
+    JANAF_URL = "http://kinetics.nist.gov/janaf/html/%s.txt"
+
+    def __init__(self):
+        """
+        We have an index file which can be used to build the url for all phases
+        on the NIST site.
+        """
+
+        # Read the index file which tells us the filenames for all the phases
+        # in the JANAF database.
+        dirname = os.path.dirname(__file__)
+        janaf_index = os.path.join(dirname, 'JANAF_index.txt')
+        self.db = pd.read_csv(janaf_index, delimiter='|')
+        # Name the columns and trim whitespace off the text fields.
+        self.db.columns = ['formula', 'name', 'phase', 'filename']
+        self.db["formula"] = self.db["formula"].map(str.strip)
+        self.db["name"] = self.db["name"].map(str.strip)
+        self.db["phase"] = self.db["phase"].map(str.strip)
+        self.db["filename"] = self.db["filename"].map(str.strip)
+
+        # Make sure that the directory for cached JANAF files exists.
+        self.JANAF_cachedir = os.path.join(dirname, 'JANAF_Cache')
+        if not os.path.exists(self.JANAF_cachedir):
+            os.mkdir(self.JANAF_cachedir)
+
+    def search(self, searchstr):
+        """
+        List all the species containing a string. Helpful for
+        interactive use of the database.
+        returns a pandas dataframe containing valid phases.
+
+        >>> db = Janafdb()
+        >>> s = db.search('Rb-')
+        >>> print(s)
+             formula           name phase filename
+        1710     Rb-  Rubidium, Ion     g   Rb-007
+        >>> s = db.search('Ti')
+        >>> print(len(s))
+        88
+        """
+
+        formulasearch = self.db['formula'].str.contains(searchstr)
+        namesearch = self.db['name'].str.contains(searchstr)
+
+        return self.db[formulasearch | namesearch]
+
+    def getphasedata(self, formula=None, name=None, phase=None, nocache=False):
+        """
+        Returns an element instance given the name of the element.
+        formula, name and phase match the respective fields in the JANAF index.
+        nocache = True means that we will always get the data from the web.
+
+        >>> db = Janafdb()
+        >>> db.getphasedata(formula='O2Ti', phase='cr')
+        Traceback (most recent call last):
+            ...
+        ValueError: There are 2 records matching this pattern.
+        >>> db.getphasedata(formula='Oxyz')
+        Traceback (most recent call last):
+            ...
+        ValueError: Valid phase types are ['cr', 'l', 'cr,l', 'g', 'ref', 'cd', 'fl', 'am', 'vit', 'mon', 'pol', 'sln', 'aq', 'sat'].
+        >>> db.getphasedata(formula='Oxyz', phase='l')
+        Traceback (most recent call last):
+            ...
+        ValueError: Did not find Oxyz, None, (l)
+        """
+
+        # Check that the phase type requested is valid.
+        if phase not in Janafdb.VALIDPHASETYPES:
+            raise ValueError("Valid phase types are " + str(
+                Janafdb.VALIDPHASETYPES) + ".")
+
+        # We can search on either an exact formula, partial text match in the
+        # name, and exact phase type.
+        formulasearch = pd.Series(np.ones(len(self.db)), dtype=bool)
+        namesearch = formulasearch.copy()
+        phasesearch = formulasearch.copy()
+        if formula is not None:
+            formulasearch = self.db['formula'] == formula
+        if name is not None:
+            namesearch = self.db['name'].str.contains(name)
+        if phase is not None:
+            phasesearch = self.db['phase'] == phase
+        searchmatch = formulasearch & namesearch & phasesearch
+
+        # Get the record (should be one record) which specifies this phase.
+        PhaseRecord = self.db[searchmatch]
+        if len(PhaseRecord) == 0:
+            raise ValueError("Did not find %s, %s, (%s)" %
+                             (formula, name, phase))
+        if len(PhaseRecord) > 1:
+            raise ValueError("There are %d records matching this pattern." %
+                             len(PhaseRecord))
+
+        # At this point we have one record.  Check if we have that file cached.
+        cachedfilename = os.path.join(
+            self.JANAF_cachedir, PhaseRecord['filename'].values[0] + '.txt')
+        if os.path.exists(cachedfilename) and not nocache:
+            # Yes it was cached, so let's read it into memory.
+            with open(cachedfilename, 'r') as f:
+                textdata = f.read()
+        else:
+            # No it was not cached so let's get it from the web.
+            response = urllib2.urlopen(Janafdb.JANAF_URL %
+                                       PhaseRecord['filename'].values[0])
+            textdata = response.read()
+
+            # And cache the data so we aren't making unnecessary trips to the
+            # web.
+            if not nocache:
+                with open(cachedfilename, 'w') as f:
+                    f.write(textdata)
+
+        # Create a phase class and return it.
+        return JanafPhase(textdata)
+
+    # def getmixturedata(self, components):
+    #     """
+    #     Creates a mixture of components given a list of tuples
+    #     containing the formula and the volume percent
+    #     """
+    #     mixture = Mixture()
+    #     for comp in components:
+    #         mixture.add(self.getelementdata(comp[0]), comp[1])
+    #
+    #     return mixture

tox.ini

@@ -5,6 +5,7 @@ envlist = py27,py34
 deps =
   nose
   numpy
+  pandas
   coverage
   setuptools
 commands =