Index-side synonym generation

A system, comprising:

• one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
• obtaining a token sequence for a resource, wherein each token in the token sequence comprises one or more characters;
• selecting a particular token from the token sequence, wherein the particular token comprises two or more characters that comprise at least one numeric portion having at least one contiguous numeric character, and at least one non-numeric portion having at least one non-numeric character;
• generating a new token for each of the at least one numeric portion that is not separately a token in the token sequence, the new token comprising the respective at least one numeric portion of the particular token without the at least one non-numeric portion of the particular token;
• and
• storing, in a search engine index, (i) data indicating the particular token and data indicating the new token as index terms for the resource, and (ii) data indicating that the particular token and the at least one numeric portion correspond to the same token in the resource.

A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:

• obtaining, at a search system, a token sequence for a resource, wherein each token in the token sequence comprises one or more characters;
• selecting, at the search system, a particular token from the token sequence, wherein the particular token comprises two or more characters that comprise at least one numeric portion having at least one contiguous numeric character, and at least one non-numeric portion having at least one non-numeric character;
• generating, at the search system and for each of the at least one numeric portion that is not separately a token in the token sequence, a new token comprising the respective at least one numeric portion of the particular token without the at least one non-numeric portion of the particular token;
• and
• storing, in a search engine index, (i) data indicating the particular token and data indicating the new token as index terms for the resource, and (ii) data indicating that the particular token and the at least one numeric portion correspond to the same token in the resource.