expression filter #57
Conversation
There was a problem hiding this comment.
This looks really nice. I made some specific code comments but overall the only thing I see missing is that I'd like to be able to apply multiple filters consecutively.
I am picturing expressions separated by a ";" (maybe) which would be parsed and applied in order to the fsd object.
| @@ -21,12 +21,19 @@ pub struct FiberFilters { | |||
| help_heading = "BAM-Options" | |||
| )] | |||
| pub bit_flag: u16, | |||
| /// Filtering expression to use for filtering records | |||
There was a problem hiding this comment.
You can extend this by adding more lines starting with ///
It would be great if you used it as a space to describe the syntax of the parser.
There was a problem hiding this comment.
Please add a description of the syntax of the parser.
| @@ -21,12 +21,19 @@ pub struct FiberFilters { | |||
| help_heading = "BAM-Options" | |||
| )] | |||
| pub bit_flag: u16, | |||
| /// Filtering expression to use for filtering records | |||
There was a problem hiding this comment.
Please add a description of the syntax of the parser.
src/utils/ftexpression.rs
Outdated
| pub fn parse_filter(filter_orig: &str) -> ParsedExpr { | ||
| let mut filter = filter_orig.to_string(); | ||
| filter.retain(|c| !c.is_whitespace()); | ||
|
|
||
| let func_name_end = filter.find('(').unwrap_or(filter.len()); | ||
| let func_name = filter[..func_name_end].trim().to_string(); | ||
|
|
||
| let gnm_feat_start = filter.find('(').unwrap_or(filter.len()) + 1; | ||
| let gnm_feat_end = filter.find(')').unwrap_or(filter.len()); | ||
| let gnm_feat = filter[gnm_feat_start..gnm_feat_end].to_string(); | ||
| if !["msp", "nuc", "m6a", "5mC"].contains(&gnm_feat.as_str()) { | ||
| eprintln!("Invalid argument for len function: {}", gnm_feat); | ||
| std::process::exit(1); | ||
| } | ||
|
|
||
| let rest = &filter[gnm_feat_end + 1..].trim(); | ||
|
|
||
| let operators = ["!=", ">=", "<=", ">", "<", "="]; | ||
| let mut operator = "".to_string(); | ||
| let mut threshold = None; | ||
| let mut range = None; | ||
|
|
||
| for &op in operators.iter() { | ||
| if let Some(pos) = rest.find(op) { | ||
| operator = op.to_string(); | ||
| let threshold_str = rest[pos + op.len()..].trim(); | ||
| if threshold_str.contains(':') { | ||
| let range_parts: Vec<&str> = threshold_str.split(':').collect(); | ||
| if range_parts.len() == 2 { | ||
| range = Some(( | ||
| range_parts[0].trim().parse::<f64>().unwrap(), | ||
| range_parts[1].trim().parse::<f64>().unwrap(), | ||
| )); | ||
| } | ||
| } else { | ||
| threshold = Some(threshold_str.parse::<f64>().unwrap()); | ||
| } | ||
| break; | ||
| } | ||
| } | ||
|
|
||
| if let Some((_, _)) = range { | ||
| if operator != "=" { | ||
| eprintln!("Range thresholds can only be used with the '=' operator."); | ||
| std::process::exit(1); | ||
| } | ||
| } | ||
|
|
||
| let threshold_value = match range { | ||
| Some((min, max)) => Threshold::Range(min, max), | ||
| None => Threshold::Single(threshold.unwrap()), | ||
| }; |
There was a problem hiding this comment.
I think it is important to be able to apply multiple filters at once in a single command such that this function returns a Vec of ParsedExp, which can then be iteritivaly applied to every fiber-seq record.
|
This looks great. Thanks! |
No description provided.