Key error "agent"

[uselessday]

When I used libinvent and run staged_learning, I get error KeyError: 'agent'
│ ❱ 377 │ │ │ terminate = optimize(package.terminator)
│ ❱ 201 │ │ results = self.scoring_function( │
│ 202 │ │ │ self.sampled.smilies, self.invalid_mask, self.duplicate_ma │
│ 203 │ │ )
I tried libinvent prior in github with "[:1]Cc2ccc1cncc(C[:2])c1c2" The github given. The error still occurs

[halx]

Hi,

many thanks for your interest in REINVENT and welcome to the community!

I do not have enough information to diagnose your problem. What are the inputs and ouputs?

Many thanks,
Hannes.

[uselessday]

Thanks for your rapid response! Please see the below toml file. Everything is not changed following the tutorial configuration setup

#This config file is designed to incentivise generation of high similarity molecules to the starting molecules
tb_logdir = "tb_logs-both"
run_type = "staged_learning"
model_type="libinvent"
uniquify_smiles = true
use_cuda = true # run on the GPU if true, on the CPU if false
json_out_config = "staged_learning.json" # write this TOML to JSON

[parameters]

LibInvent: find R-groups for the given scaffolds

prior_file = "ibinvent.prior"
agent_file = "libinvent.prior"
smiles_file = "both.smi" # 1 scaffold per line with attachment points
#both.smi:[:1]Cc2ccc1cncc(C[:2])c1c2

num_smiles = 10000 # number of SMILES to be sampled, 1 per input SMILES
unique_molecules = true # if true remove all duplicatesd canonicalize smiles
randomize_smiles = true # if true shuffle atoms in SMILES randomly

summary_csv_prefix = "staged"
use_checkpoint = false # if true read diversity filter from agent_file
purge_memories = false # if true purge all diversity filter memories after each stage

[learning_strategy]

type = "dap" # dap: only one supported
sigma = 128 # sigma of the RL reward function
rate = 0.0001 # for torch.optim

[diversity_filter] # optional, comment section out or remove if unneeded

NOTE: also memorizes all seen SMILES

type = "PenalizeSameSmiles"
#type = "IdenticalMurckoScaffold" # IdenticalTopologicalScaffold,

ScaffoldSimilarity, PenalizeSameSmiles

bucket_size = 10 # memory size in number of compounds
minscore = 0.4
minsimilarity = 0.4
penalty_multiplier = 0.5

Stage 1

Note that stages must always be a list i.e. double brackets

[[stage]]

chkpt_file = "libinvent.prior"

termination = "simple" # termination criterion fot this stage
max_score = 0.95 # terminate if this total score is exceeded
min_steps = 100 # run for at least this number of steps
max_steps = 1000 # terminate entire run when exceeded

[stage.scoring]
type = "geometric_mean" # aggregation function

[[stage.scoring.component]]
[stage.scoring.component.custom_alerts]
[[stage.scoring.component.custom_alerts.endpoint]]
name = "Unwanted SMARTS" # user chosen name for output
weight = 0.2 # weight to fine-tune the relevance of this component
params.smarts = [
"[;r8]",
"[;r9]",
"[;r10]",
"[;r11]",
"[;r12]",
"[;r13]",
"[;r14]",
"[;r15]",
"[;r16]",
"[;r17]",
"[#8][#8]",
"[#6;+]",
"[#16][#16]",
"[#7;!n][S;!$(S(=O)=O)]",
"[#7;!n][#7;!n]",
"C#C",
"C(=[O,S])[O,S]",
"[#7;!n][C;!$(C(=[O,N])[N,O])][#16;!s]",
"[#7;!n][C;!$(C(=[O,N])[N,O])][#7;!n]",
"[#7;!n][C;!$(C(=[O,N])[N,O])][#8;!o]",
"[#8;!o][C;!$(C(=[O,N])[N,O])][#16;!s]",
"[#8;!o][C;!$(C(=[O,N])[N,O])][#8;!o]",
"[#16;!s][C;!$(C(=[O,N])[N,O])][#16;!s]",
]

[[stage.scoring.component]]
[stage.scoring.component.MolecularWeight]
[[stage.scoring.component.MolecularWeight.endpoint]]
name = "Molecular weight"
weight = 0.8
transform.type = "double_sigmoid"
transform.high = 500.0
transform.low = 350.0
transform.coef_div = 500.0
transform.coef_si = 20.0
transform.coef_se = 20.0

[halx]

Your input configuration file looks rather inconsistent to me and suggests that you are not using a recent version of the software. It would be good if you did that such that you have a sensible setup and we can take it from there. Many thanks.