rdkit-repr

Name: rdkit-repr
Author: jinzhezenggroup

113stars23forks

How to install rdkit-repr?

jinzhezenggroup/computational-chemistry-agent-skills/rdkit-repr

$npx -y skills add jinzhezenggroup/computational-chemistry-agent-skills --skill rdkit-repr

Installs into the current project.

Summary

A standardized CLI wrapper for RDKit molecular featurization workflows that handles physicochemical descriptor computation (outputs .csv) and molecular fingerprint extraction (outputs .npy or .csv), with built-in SMILES validation. USE WHEN you need to compute RDKit molecular des

Files · 1

View on GitHub

SKILL.md

1	# RDKit Molecular Featurization
2
3	This skill provides practical command patterns for RDKit descriptor and fingerprint extraction
4	using the standardized CLI wrapper: `<skill_path>/scripts/rdkit_helper.py`.
5
6	Key behaviors (important for Agents):
7
8	- The script prints environment detection (Python/RDKit/NumPy/Pandas) by default.
9	- Bad/illegal SMILES are skipped and logged to `*.skipped.csv` (no crash).
10	- Each run ends by printing absolute output paths like:
11	- `[RESULT] desc_csv=/abs/path.csv`
12	- `[RESULT] fp_npy=/abs/path.npy`
13	- `[RESULT] fp_csv=/abs/path.csv`
14
15	## Quick Start
16
17	Check CLI help:
18
19	```bash
20	uv run <skill_path>/scripts/rdkit_helper.py --help
21	```
22
23	Check subcommand help:
24
25	```bash
26	uv run <skill_path>/scripts/rdkit_helper.py desc --help
27	uv run <skill_path>/scripts/rdkit_helper.py fp --help
28	uv run <skill_path>/scripts/rdkit_helper.py list-desc --help
29	```
30
31	Disable environment printing (optional):
32
33	```bash
34	uv run <skill_path>/scripts/rdkit_helper.py --no-env desc --smiles "CCO" --output out.csv
35	```
36
37	## Core Tasks
38
39	### 1) Compute physicochemical descriptors → .csv
40
41	Single SMILES (default preset: `physchem`, 25 descriptors):
42
43	```bash
44	uv run <skill_path>/scripts/rdkit_helper.py desc \
45	--smiles "CCO" \
46	--output /tmp/CCO.desc.csv
47	```
48
49	From CSV (default SMILES column is `smiles`):
50
51	```bash
52	uv run <skill_path>/scripts/rdkit_helper.py desc \
53	--file data.csv \
54	--smiles-col smiles \
55	--output data.desc.csv
56	```
57
58	From SMI:
59
60	```bash
61	uv run <skill_path>/scripts/rdkit_helper.py desc \
62	--file molecules.smi \
63	--output molecules.desc.csv
64	```
65
66	Choose a descriptor preset:
67
68	```bash
69	# Lipinski drug-likeness (6 descriptors: MolWt, MolLogP, NumHDonors, ...)
70	uv run <skill_path>/scripts/rdkit_helper.py desc \
71	--file data.csv --preset lipinski --output data.lipinski.csv
72
73	# Extended physicochemical (25 descriptors, default)
74	uv run <skill_path>/scripts/rdkit_helper.py desc \
75	--file data.csv --preset physchem --output data.physchem.csv
76
77	# Topological / graph indices (56 descriptors: BalabanJ, BertzCT, Chi, PEOE_VSA, ...)
78	uv run <skill_path>/scripts/rdkit_helper.py desc \
79	--file data.csv --preset topological --output data.topo.csv
80
81	# All RDKit descriptors (~200 descriptors)
82	uv run <skill_path>/scripts/rdkit_helper.py desc \
83	--file data.csv --preset all --output data.all_desc.csv
84	```
85
86	Select specific descriptors (overrides `--preset`):
87
88	```bash
89	uv run <skill_path>/scripts/rdkit_helper.py desc \
90	--file data.csv \
91	--descriptors "MolWt,MolLogP,TPSA,NumHDonors,NumHAcceptors" \
92	--output data.custom.csv
93	```
94
95	Suppress merging back original CSV columns (output only smiles + descriptors):
96
97	```bash
98	uv run <skill_path>/scripts/rdkit_helper.py desc \
99	--file data.csv --preset physchem --no-merge --output data.desc_only.csv
100	```
101
102	______________________________________________________________________
103
104	### 2) Compute molecular fingerprints → .npy or .csv
105
106	Available fingerprint types:
107
108	\| Type \| Description \| Default bits \|
109	\| --------------- \| ---------------------------------------------------------- \| ------------ \|
110	\| `morgan2` \| Morgan circular FP radius 2 (ECFP4-like), bit vector \| 2048 \|
111	\| `morgan3` \| Morgan circular FP radius 3 (ECFP6-like), bit vector \| 2048 \|
112	\| `morgan2_count` \| Morgan radius-2 count vector \| 2048 \|
113	\| `rdkit` \| RDKit path-based FP, bit vector \| 2048 \|
114	\| `maccs` \| MACCS 167 structural keys (bit vector, `--nbits` ignored) \| 167 \|
115	\| `topological` \| Topological torsion FP (count vector, hashed to `--nbits`) \| 2048 \|
116	\| `atompair` \| Atom-pair FP (count vector, hashed to `--nbits`) \| 2048 \|
117	\| `layered` \| Layered substructure FP, bit vector \| 2048 \|
118	\| `pattern` \| SMARTS pattern FP, bit vector \| 2048 \|
119
120	Single SMILES, output as NumPy array (.npy):
121
122	```bash
123	uv run <skill_path>/scripts/rdkit_helper.py

People who liked this also viewed

6 picks

Type

rdkit-repr

113stars23forks

View on GitHub

How to install rdkit-repr?

jinzhezenggroup/computational-chemistry-agent-skills/rdkit-repr

$npx -y skills add jinzhezenggroup/computational-chemistry-agent-skills --skill rdkit-repr

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# RDKit Molecular Featurization
2
3	This skill provides practical command patterns for RDKit descriptor and fingerprint extraction
4	using the standardized CLI wrapper: `<skill_path>/scripts/rdkit_helper.py`.
5
6	Key behaviors (important for Agents):
7
8	- The script prints environment detection (Python/RDKit/NumPy/Pandas) by default.
9	- Bad/illegal SMILES are skipped and logged to `*.skipped.csv` (no crash).
10	- Each run ends by printing absolute output paths like:
11	- `[RESULT] desc_csv=/abs/path.csv`
12	- `[RESULT] fp_npy=/abs/path.npy`
13	- `[RESULT] fp_csv=/abs/path.csv`
14
15	## Quick Start
16
17	Check CLI help:
18
19	```bash
20	uv run <skill_path>/scripts/rdkit_helper.py --help
21	```
22
23	Check subcommand help:
24
25	```bash
26	uv run <skill_path>/scripts/rdkit_helper.py desc --help
27	uv run <skill_path>/scripts/rdkit_helper.py fp --help
28	uv run <skill_path>/scripts/rdkit_helper.py list-desc --help
29	```
30
31	Disable environment printing (optional):
32
33	```bash
34	uv run <skill_path>/scripts/rdkit_helper.py --no-env desc --smiles "CCO" --output out.csv
35	```
36
37	## Core Tasks
38
39	### 1) Compute physicochemical descriptors → .csv
40
41	Single SMILES (default preset: `physchem`, 25 descriptors):
42
43	```bash
44	uv run <skill_path>/scripts/rdkit_helper.py desc \
45	--smiles "CCO" \
46	--output /tmp/CCO.desc.csv
47	```
48
49	From CSV (default SMILES column is `smiles`):
50
51	```bash
52	uv run <skill_path>/scripts/rdkit_helper.py desc \
53	--file data.csv \
54	--smiles-col smiles \
55	--output data.desc.csv
56	```
57
58	From SMI:
59
60	```bash
61	uv run <skill_path>/scripts/rdkit_helper.py desc \
62	--file molecules.smi \
63	--output molecules.desc.csv
64	```
65
66	Choose a descriptor preset:
67
68	```bash
69	# Lipinski drug-likeness (6 descriptors: MolWt, MolLogP, NumHDonors, ...)
70	uv run <skill_path>/scripts/rdkit_helper.py desc \
71	--file data.csv --preset lipinski --output data.lipinski.csv
72
73	# Extended physicochemical (25 descriptors, default)
74	uv run <skill_path>/scripts/rdkit_helper.py desc \
75	--file data.csv --preset physchem --output data.physchem.csv
76
77	# Topological / graph indices (56 descriptors: BalabanJ, BertzCT, Chi, PEOE_VSA, ...)
78	uv run <skill_path>/scripts/rdkit_helper.py desc \
79	--file data.csv --preset topological --output data.topo.csv
80
81	# All RDKit descriptors (~200 descriptors)
82	uv run <skill_path>/scripts/rdkit_helper.py desc \
83	--file data.csv --preset all --output data.all_desc.csv
84	```
85
86	Select specific descriptors (overrides `--preset`):
87
88	```bash
89	uv run <skill_path>/scripts/rdkit_helper.py desc \
90	--file data.csv \
91	--descriptors "MolWt,MolLogP,TPSA,NumHDonors,NumHAcceptors" \
92	--output data.custom.csv
93	```
94
95	Suppress merging back original CSV columns (output only smiles + descriptors):
96
97	```bash
98	uv run <skill_path>/scripts/rdkit_helper.py desc \
99	--file data.csv --preset physchem --no-merge --output data.desc_only.csv
100	```
101
102	______________________________________________________________________
103
104	### 2) Compute molecular fingerprints → .npy or .csv
105
106	Available fingerprint types:
107
108	\| Type \| Description \| Default bits \|
109	\| --------------- \| ---------------------------------------------------------- \| ------------ \|
110	\| `morgan2` \| Morgan circular FP radius 2 (ECFP4-like), bit vector \| 2048 \|
111	\| `morgan3` \| Morgan circular FP radius 3 (ECFP6-like), bit vector \| 2048 \|
112	\| `morgan2_count` \| Morgan radius-2 count vector \| 2048 \|
113	\| `rdkit` \| RDKit path-based FP, bit vector \| 2048 \|
114	\| `maccs` \| MACCS 167 structural keys (bit vector, `--nbits` ignored) \| 167 \|
115	\| `topological` \| Topological torsion FP (count vector, hashed to `--nbits`) \| 2048 \|
116	\| `atompair` \| Atom-pair FP (count vector, hashed to `--nbits`) \| 2048 \|
117	\| `layered` \| Layered substructure FP, bit vector \| 2048 \|
118	\| `pattern` \| SMARTS pattern FP, bit vector \| 2048 \|
119
120	Single SMILES, output as NumPy array (.npy):
121
122	```bash
123	uv run <skill_path>/scripts/rdkit_helper.py

People who liked this also viewed

6 picks

Type