背景
我从互联网上的一个来源获取了一个表(请参阅Mordred Molecular Descriptors)机器学习项目。
我用来获取该表的代码如下所示：
import requests
import pandas as pd
from bs4 import BeautifulSoup
# Fetch the HTML content of the webpage
url = "https://mordred-descriptor.github.io/documentation/master/descriptors.html"
html = requests.get(url).content
# Parse the HTML content using BeautifulSoup
soup = BeautifulSoup(html, 'html.parser')
# Find the table element in the HTML
table = soup.find('table')
# Convert the table into a Pandas dataframe
df = pd.read_html(str(table))[0]
# Print the resulting dataframe
df.drop(['#', 'constructor', 'dim', 'description'], axis=1)
在Python3中运行以上代码后，我生成了这个数据帧。
现在我想把"；名称"；列中的相应模块；模块"；列
问题是提取的表已经是透视表；模块"；列由NaN值填充。理想情况下，我想生成一个字典，其中包含键、模块和值，以及分组名称的列表。
示例：dict_df = {'ABCIndex': ['ABC','ABCGG'], 'AcidBase': ['nAcid', 'nBase'], ..., 'ZagrebIndex': ['Zagreb1', 'Zagreb2', 'mZagreb1', 'mZagreb2']}
我曾尝试在Pandas中使用.groupby()按模块将名称分组，但NaN值被省略了，字典值只剩下一个名称列表；模块不是NaN值的行的名称。
感谢您的时间和帮助。
IIUC，像这样？使用ffill，然后使用groupby，使用agg和list。
df.groupby(df['module'].ffill())['name'].agg(list)
输出：
module
ABCIndex [ABC, ABCGG]
AcidBase [nAcid, nBase]
AdjacencyMatrix [SpAbs_A, SpMax_A, SpDiam_A, SpAD_A, SpMAD_A, ...
Aromatic [nAromAtom, nAromBond]
AtomCount [nAtom, nHeavyAtom, nSpiro, nBridgehead, nHete...
Autocorrelation [ATS0dv, ATS1dv, ATS2dv, ATS3dv, ATS4dv, ATS5d...
BCUT [BCUTc-1h, BCUTc-1l, BCUTdv-1h, BCUTdv-1l, BCU...
BalabanJ [BalabanJ]
BaryszMatrix [SpAbs_DzZ, SpMax_DzZ, SpDiam_DzZ, SpAD_DzZ, S...
BertzCT [BertzCT]
BondCount [nBonds, nBondsO, nBondsS, nBondsD, nBondsT, n...
CPSA [PNSA1, PNSA2, PNSA3, PNSA4, PNSA5, PPSA1, PPS...
CarbonTypes [C1SP1, C2SP1, C1SP2, C2SP2, C3SP2, C1SP3, C2S...
Chi [Xch-3d, Xch-4d, Xch-5d, Xch-6d, Xch-7d, Xch-3...
Constitutional [SZ, Sm, Sv, Sse, Spe, Sare, Sp, Si, MZ, Mm, M...
DetourMatrix [SpAbs_Dt, SpMax_Dt, SpDiam_Dt, SpAD_Dt, SpMAD...
DistanceMatrix [SpAbs_D, SpMax_D, SpDiam_D, SpAD_D, SpMAD_D, ...
EState [NsLi, NssBe, NssssBe, NssBH, NsssB, NssssB, N...
EccentricConnectivityIndex [ECIndex]
ExtendedTopochemicalAtom [ETA_alpha, AETA_alpha, ETA_shape_p, ETA_shape...
FragmentComplexity [fragCpx]
Framework [fMF]
GeometricalIndex [GeomDiameter, GeomRadius, GeomShapeIndex, Geo...
GravitationalIndex [GRAV, GRAVH, GRAVp, GRAVHp]
HydrogenBond [nHBAcc, nHBDon]
InformationContent [IC0, IC1, IC2, IC3, IC4, IC5, TIC0, TIC1, TIC...
KappaShapeIndex [Kier1, Kier2, Kier3]
Lipinski [Lipinski, GhoseFilter]
LogS [FilterItLogS]
McGowanVolume [VMcGowan]
MoRSE [Mor01, Mor02, Mor03, Mor04, Mor05, Mor06, Mor...
MoeType [LabuteASA, PEOE_VSA1, PEOE_VSA2, PEOE_VSA3, P...
MolecularDistanceEdge [MDEC-11, MDEC-12, MDEC-13, MDEC-14, MDEC-22, ...
MolecularId [MID, AMID, MID_h, AMID_h, MID_C, AMID_C, MID_...
MomentOfInertia [MOMI-X, MOMI-Y, MOMI-Z]
PBF [PBF]
PathCount [MPC2, MPC3, MPC4, MPC5, MPC6, MPC7, MPC8, MPC...
Polarizability [apol, bpol]
RingCount [nRing, n3Ring, n4Ring, n5Ring, n6Ring, n7Ring...
RotatableBond [nRot, RotRatio]
SLogP [SLogP, SMR]
TopoPSA [TopoPSA(NO), TopoPSA]
TopologicalCharge [GGI1, GGI2, GGI3, GGI4, GGI5, GGI6, GGI7, GGI...
TopologicalIndex [Diameter, Radius, TopoShapeIndex, PetitjeanIn...
VdwVolumeABC [Vabc]
VertexAdjacencyInformation [VAdjMat]
WalkCount [MWC01, MWC02, MWC03, MWC04, MWC05, MWC06, MWC...
Weight [MW, AMW]
WienerIndex [WPath, WPol]
ZagrebIndex [Zagreb1, Zagreb2, mZagreb1, mZagreb2]
Name: name, dtype: object