ivanchernyshov / nistchempy Goto Github PK

For aspirin (ID = C50782) instead of

['Benzoic acid, 2-(acetyloxy)-', 'Salicylic acid acetate']

gets:

['Benzoic', 'acid,', '2-(acetyloxy)-', 'Salicylic', 'acid', 'acetate']

Transform the tutorial into the Sphinx/ReadTheDocs documentation

NistChemPy 0.1.1 processes mol2d, mol3d, ir, ms, and uvvis only. It would be nice to add other properties so that they will have specified short keys in the data_refs dictionary.

Use the sitemap to extract all compounds from NIST Chemistry WebBook as table with available properties as columns

Change the code so that in the absence of the necessary python libraries some functionality does not work, e.g. no parsed search without pandas, no substructure search without RDKit, and so on.

The final requirements must be requests, BeautifulSoup, and importlib-resources due to the problems with the python version.

NistChemPy 0.1.1, "Reaction thermochemistry data" is not in the keys if there are more than 50 reactions:

>>> import nistchempy as nist
>>> x = nist.Compound('C13968086')
>>> x.data_refs
{'mol2d': 'https://webbook.nist.gov/cgi/cbook.cgi?Str2File=C13968086',
 'mol3d': 'https://webbook.nist.gov/cgi/cbook.cgi?Str3File=C13968086',
 'Gas phase thermochemistry data': 'https://webbook.nist.gov/cgi/cbook.cgi?ID=C13968086&Units=SI&Mask=1#Thermo-Gas',
 'reactions 1 to 50': 'https://webbook.nist.gov/cgi/cbook.cgi?ID=C13968086&Units=SI&Mask=8#Thermo-React',
 'reactions 51 to 51': 'https://webbook.nist.gov/cgi/cbook.cgi?ID=C13968086&Units=SI&Mask=8&StartReac=50#Thermo-React',
 'Ion clustering data': 'https://webbook.nist.gov/cgi/cbook.cgi?ID=C13968086&Units=SI&Mask=40#Ion-Cluster',
 'Vibrational and/or electronic energy levels': 'https://webbook.nist.gov/cgi/cbook.cgi?ID=C13968086&Units=SI&Mask=800#Electronic-Spec'}

Instead of just returning dataframe with info on compounds, make an object for the search using this dataframe. For the purpose:

1. Reparse synonyms from the NIST database and update the dadaframe
2. Implement the name search
3. Clear chemical formulas and implement the formula search
4. Use RDKit to implement the substructure search

Add the same interface as for IR, UV, and MS spectra

Search by InChI always gives compound-like HTML page even if WebBook does not contain this compound:

>>> import nistchempy as nist
>>> S = nist.Search()
>>> S.find_compounds('InChI=1S/Al.H/i;1+1')

Traceback (most recent call last):

  File "C:\Users\Grim\AppData\Local\Temp/ipykernel_9400/3310466150.py", line 1, in <module>
    S.find_compounds(inchis[0], 'inchi')

  File "C:\Programming\Anaconda3\envs\csd\lib\site-packages\nistchempy.py", line 427, in find_compounds
    refs = soup.find('ol').findChildren('a', href = re.compile(self._COMP_ID))

AttributeError: 'NoneType' object has no attribute 'findChildren'

Searches with high requirements to compounds (e.g. to have all possible properties) return a list of compounds which possess just some of the required properties. Possibly there are a way to detect if the search was unsuccessful ("3FFF" mask?).