zet

Zettelkasten Search Approaches

The way I manage a Zettelkasten-ish repo is as simple as possible relying on the least amount of technology, in this case minimal bash shell scripting (with no assumption that jq is available). Indeed, traditional, delimited text files remain the fastest and easiest to read structured data file format.

What would I want to query?

We have to have a search engine design that allows the person doing the search to have the most power. But what would those search queries look like? Here are some examples:

• Most recently created zettels
• Most recently updated zettels
• Single keyword anywhere in the text of any zettel
• Any of multiple keywords in text of any zettel
• All of multiple keywords in text of any zettel
• Regular expression match in the text of a any zettel
• Any of the above, but just for titles, or headings, or bodies,
• Any of the above, but only first first paragraph
• Any of the above, but only within code block
• Any of the above, but only within verbatim block
• Any of the above, but only within quoted block

Searchable Block Scope

Here’s a description of the different types of blocks that make up any ZettelMark file. Scopes can be added to queries by combining their prefix identifiers:

• Title (t)
• Heading (h)
• Heading1 (h1) .. Heading6 (h6)
• Paragraph (p)
• Code (c)
• Verbatim (v)
• Quote (q)
• Table (tbl)
• List (l)
• OrderedList (ol)
• UnorderedList (ul)
• Tags (tag)
• Video (vid)

Zettel Query Language Syntax

The default query language is simply PCRE, which is activated with -P on modern grep. Training users in multiple variations of querying their data is against their best interests. Perhaps, however, we could have a configuration variable that sets the query language to use:

• Keyword
• Extended glob
• ERE
• PCRE

Will just support PCRE for now.

Case sensitivity is off by default.

Combining Query Scope

Scopes can be combined together with one another (joined with an underscore, which is query string friendly) and then finally tied to a specific query. The order of the scope prefixes determines priority when ranking and ordering results. To reserve, use exclamation point instead of equals sign.

Query	Description
l=foo	Any list item contains the keyword foo
l!foo	Any list item that does not contain the keyword foo
ul_ol=foo	Any list item contains the keyword foo, unordered first
tag=#foo\b	Tags line contains foo hashtag
h=fo+	Any heading with matching fo+ regular expression (PCRE)

Of course, these will all be URL encoded when sent over a web request.

tag=#foo
tag=%23foo
h=fo%2b

When multiple queries are combined each should be separated by a single space. Usually this means each query will be a separate argument. When multiple queries are passed the first query matches will have priority in the results.

Optional Tags Line

If the last line of the zettel begins with four or more spaces followed by an octothorpe (#) then it will be interpreted as a Tags line/block.

#tips #zettekasten #streaming

This site is open source. Improve this page.