Conversation
* Added catechol dataset pbtxt * fixed yields * changes from review * actual review changes, removal of template
Added enhanced name and description to the dataset, and corrected the 'reactino' typo using search/replace.
|
Change summary:
|
|
@jpfolch your ORD dataset ID has now been created. It is ord_dataset-1ec2807f02fa4beda27d2a81b86eb843 and your dataset will be found here in the ord-data repo: I need to give the automated checks a kick to get the count to finish properly, but it should flow through to the public database from here. |
|
Change summary:
|
|
Change summary:
|
|
Thanks @bdeadman, all ready and looking good from my end. |
|
@skearnes - this dataset was submitted by the authors, reviewed by me, and now needs approval to finalise the PR. |
|
Change summary:
|
|
Approved. This is a well-structured transient flow dataset with good provenance, consistent compound identifiers, and properly populated flow conditions. Automated validation: All 259 warnings flag yield values as "possible fractions (0–1 scale)" and are false positives. The affected reactions are early-conversion datapoints where Product 2 and Product 3 are genuinely at sub-1% HPLC area, with starting material at 87–94%. The validator has no context to distinguish a legitimate low yield from a scaling error. Notes for future revision:
Completeness: Good |
|
Change summary:
|
Populate ReactionProvenance.doi with 10.48550/arXiv.2506.07619 across all 1,227 reactions (previously only set in publication_url). Set is_desired_product = True for Product 2 and Product 3 in all reactions.
|
Updated dataset uploaded addressing the two points raised in my earlier post-review review:
|
|
Change summary:
|
|
Change summary:
|
|
@jpfolch - sorry for the delays in getting this online. Its up now and can be found at https://open-reaction-database.org/dataset/ord_dataset-1ec2807f02fa4beda27d2a81b86eb843. I've set up a LinkedIn post about these datasets (and another separate release) to go out at 10.30 BST today. |
* Backfill parquet versions for 3 datasets added after #241 The submissions merged after the parquet conversion (#235 catechol transient flow, #238 decarboxylative olefination x2) only had .pb.gz. Generated 1:1 parquet siblings with scripts/convert_to_parquet.py (all singletons; no de-shard merge applies). Reaction counts match the originals (1227/120/136) and validate_dataset.py passes. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Document Parquet conversion and LFS push setup for maintainers Add README maintainer notes for backfilling Parquet with convert_to_parquet.py and for routing new-LFS-object pushes to GitHub (since .lfsconfig points reads at the HF mirror and has no pushurl). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Use two-space inline comments in README code blocks Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Backfill parquet version for Cernak C–N coupling dataset (#243) The Cernak 50,688-reaction Pd/Ni/Cu C–N coupling dataset (#243) was merged after the parquet conversion (#241), so it only had .pb.gz. Generated its 1:1 parquet sibling with scripts/convert_to_parquet.py (singleton path; the name matches no MERGE_SPEC, so it keeps its existing dataset_id). Reaction count matches the original (50688) and the parquet round-trips. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
3 participants
dataset name:
Transient flow exploration of the effect of temperature, residence time and binary solvent ratios on a uni-molecular allyl substituted catechol reaction.
dataset description:
1,227 datapoints were collected in a series of transient flow ramps involving temperature [175 - 225 C], residence time [2 - 15 minutes] and binary solvent ratios of 24 solvents. HPLC yields reported for starting material and 2 products of the uni-molecular rearrangement of o-Bis(allyloxy)benzene. Reactions performed by SOLVE Chemistry in London (solvechemistry.com) and published in https://doi.org/10.48550/arXiv.2506.07619.
Dataset reviewed by me in #233 and authors made the required corrections already. I corrected a typo, and enhanced the dataset name and description in the latest commit on this branch.