From 0adb490ae2ba8f961bac559a981f89d6d264af60 Mon Sep 17 00:00:00 2001 From: Bryan Newbold Date: Mon, 18 Jan 2021 19:52:37 -0800 Subject: initial notes on crude query parsing --- proposals/2021_crude_query_parse.md | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) create mode 100644 proposals/2021_crude_query_parse.md (limited to 'proposals') diff --git a/proposals/2021_crude_query_parse.md b/proposals/2021_crude_query_parse.md new file mode 100644 index 0000000..2a7663b --- /dev/null +++ b/proposals/2021_crude_query_parse.md @@ -0,0 +1,18 @@ + + +Thinking of simple ways to reduce query parse errors and handle more queries as +expected. In particular: + +- handle slashes in query tokens (eg, "N/A" without quotes) +- handle semi-colons in queries, when they are not intended as filters +- if query "looks like" a raw citation string, detect that and do citation + parsing in to a structured format, then do a query or fuzzy lookup from there + + +## Questions/Thoughts + +Should we detect title lookups in addition to full citation lookups? Probably +too complicated. + +Do we have a static list of colon-prefixes, or load from the schema mapping +file itself? -- cgit v1.2.3