chemcaption smarts (#456)

data/tabular/chem_caption_smarts/meta.yaml

@@ -0,0 +1,41 @@
+---
+name: chem_caption_smarts
+description: |-
+    This dataset contains the count of substructures in molecules
+targets:
+    - id: smarts
+      type: text
+      description: substructure smarts
+      names:
+          - noun: SMARTS
+          - noun: SMiles ARbitrary Target Specification (SMARTS)
+    - id: completion
+      type: categorical
+      description: number of matches
+    - id: completion_labels
+      type: text
+      description: name of the substructure
+identifiers:
+    - id: representation
+      type: text
+      description: representation
+    - id: representation_type
+      type: text
+      description: representation type
+license: CC BY 4.0
+links:
+    - url: https://github.com/lamalab-org/chem-caption
+      description: Original codebase used to generate this dataset
+templates:
+    - |-
+      Question: {#How many times|How often!} does the {#molecule|chemical|compound|chemical structure!} with {representation_type#} {representation#} contain the substructure with the {smarts__names__noun} {#smarts#}?
+      Answer: {completion#}
+    - |-
+      Question: {#How many times|How often!} does the {#molecule|chemical|compound|chemical structure!} with {representation_type#} {representation#} contain a {completion#} substructure?
+      Answer: {smarts__names__noun} {#smarts#}
+    - |-
+      User: {#I want to|I have to|I must|I would like to!} know {#how many times|how often!} the {#molecule|chemical|compound|chemical structure!} with {representation_type#} {representation#} contains the substructure with the {smarts__names__noun} {#smarts#}.
+      Assistant: The {#molecule|chemical|compound|chemical structure!} with {representation_type#} {representation#} contains the substructure with the {smarts__names__noun} {#smarts#} {completion#} times.
+    - |-
+      User: {#I want to|I have to|I must|I would like to!} know how many times the {#molecule|chemical|compound|chemical structure!} with {representation_type#} {representation#} contains a {completion#} substructure.
+      Assistant: The {#molecule|chemical|compound|chemical structure!} contains the substructure with the {smarts__names__noun} {#smarts#} {completion#} times.