cleanup csq stuff

scripts/csq.lua

@@ -76,3 +76,72 @@ print(c.Allele)
 print(c.AF)
 
 --]]
+
+function parse_description(description)
+    -- Remove quotes and spaces
+    description = description:gsub('"', ''):gsub('%s+', '')
+
+    -- Determine the correct split point
+    local format_str = nil
+    if description:find('Format:') then
+        format_str = description:match('Format:(.+)')
+    else
+        format_str = description:match(':\'(.+)')
+    end
+
+    if format_str then
+        -- Split by | delimiter and store in a table
+        local result = {}
+        for value in string.gmatch(format_str, '([^|]+)') do
+            -- Remove any trailing characters like ')' or ']'
+            value = value:gsub('[%)>\'%]%[]', '')
+            -- test if value startswith 'Effect(' and remove it if that's the case.
+            if value:find('Effect%(') then
+                value = value:gsub('^\'?Effect%(', '')
+            end
+
+            table.insert(result, value)
+        end
+        return result
+    else
+        return nil, "Format part not found in description."
+    end
+end
+
+--[[
+-- Example usage
+local input1 =
+'#INFO=<ID=CSQ,Number=.,Type=String,Description="Consequence type as predicted by VEP. Format: Consequence|Codons|Amino_acids|Gene|SYMBOL|Feature|EXON|PolyPhen|SIFT|Protein_position|BIOTYPE">'
+local input2 =
+'##INFO=<ID=EFF,Number=.,Type=String,Description="Predicted effects for this variant.Format: \'Effect ( Effect_Impact | Functional_Class | Codon_Change | Amino_Acid_change| Amino_Acid_length | Gene_Name | Gene_BioType | Coding | Transcript | Exon [ | ERRORS | WARNINGS ] )\'">'
+local input3 =
+'##INFO=<ID=ANN,Number=.,Type=String,Description="Functional annotations: \'Allele | Annotation | Annotation_Impact | Gene_Name | Gene_ID | Feature_Type | Feature_ID | Transcript_BioType | Rank | HGVS.c | HGVS.p | cDNA.pos / cDNA.length | CDS.pos / CDS.length | AA.pos / AA.length | Distance | ERRORS / WARNINGS / INFO\'">'
+
+local parsed_table1 = parse_description(input1)
+local parsed_table2 = parse_description(input2)
+local parsed_table3 = parse_description(input3)
+
+print("Parsed Table 1:")
+for i, v in ipairs(parsed_table1) do
+    print(i, v)
+end
+
+print("\nParsed Table 2:")
+for i, v in ipairs(parsed_table2) do
+    print(i, v)
+end
+
+print("\nParsed Table 3:")
+for i, v in ipairs(parsed_table3) do
+    print(i, v)
+end
+
+
+desc =
+'Functional annotations: \'Allele | Annotation | Annotation_Impact | Gene_Name | Gene_ID | Feature_Type | Feature_ID | Transcript_BioType | Rank | HGVS.c | HGVS.p | cDNA.pos / cDNA.length | CDS.pos / CDS.length | AA.pos / AA.length | Distance | ERRORS / WARNINGS / INFO\'"'
+
+for i, v in ipairs(parse_description(desc)) do
+    print(i, v)
+end
+
+--]]

scripts/pre.lua

@@ -1,4 +1,5 @@
-header:add_info({ID="af_copy", Number=1, Description="adding a single field", Type="Float"})
+header:add_info({ ID = "af_copy", Number = 1, Description = "adding a single field", Type = "Float" })
 --print(header:info_get("af_copy").Description)
 
 header_desc = header:info_get("vep").Description
+desc = parse_description(header_desc)

scripts/t.sh

@@ -1 +1,6 @@
-./target/debug/vcfexpr filter -e "print(CSQ.new(variant:info('vep', 0), desc).Allele); return variant.id == 'rs2124717267'"  -o var.bcf gnomad.genomes.v4.0.sites.chrY.vcf.bgz -p scripts/pre.lua -s 'af_copy=return variant:info("AF", 0)' -l scripts/csq.lua -l scripts/csq_header.lua   -l  scripts/h.lua
+# load the csq code (csq.lua), then add the af_copy field and define `desc` from the header (pre.lua)
+./target/debug/vcfexpr filter -e "csq = CSQ.new(variant:info('vep', 0), desc); return csq.IMPACT == 'HIGH'" \
+        -o var.bcf gnomad.genomes.v4.0.sites.chrY.vcf.bgz \
+        -p scripts/csq.lua \
+        -p scripts/pre.lua \
+        -s 'af_copy=return variant:info("AF", 0)' 

src/main.rs

@@ -36,12 +36,15 @@ pub enum Commands {
         template: Option<String>,
 
         /// File(s) containing lua code to load. May contain functions that will be called by the expressions.
+        /// Prefer to use `--lua_prelude` for code that should be run once before any variants are processed.
+        /// This option may be deprecated.
         #[arg(short, long)]
         lua: Vec<String>,
 
-        /// File containing lua code to run once before any variants are processed.
+        /// File(s) containing lua code to run once before any variants are processed.
+        /// `header` is available here to access or modify the header.
         #[arg(short = 'p', long)]
-        lua_prelude: Option<String>,
+        lua_prelude: Vec<String>,
 
         /// Optional output file. Default is stdout.
         #[arg(short, long)]
@@ -55,7 +58,7 @@ fn filter_main(
     set_expression: Vec<String>,
     template: Option<String>,
     lua_code: Vec<String>,
-    lua_prelude: Option<String>,
+    lua_prelude: Vec<String>,
     output: Option<String>,
 ) -> Result<(), Box<dyn std::error::Error>> {
     env_logger::init();

src/vcfexpr.rs

@@ -136,7 +136,7 @@ impl<'lua> VCFExpr<'lua> {
         expression: Vec<String>,
         set_expression: Vec<String>,
         template: Option<String>,
-        lua_prelude: Option<String>,
+        lua_prelude: Vec<String>,
         output: Option<String>,
     ) -> Result<Self, Box<dyn std::error::Error>> {
         lua.load(crate::pprint::PPRINT).set_name("pprint").exec()?;
@@ -167,13 +167,14 @@ impl<'lua> VCFExpr<'lua> {
             rust_htslib::htslib::bcf_hdr_dup(reader.header().inner)
         });
 
-        if let Some(lua_code) = lua_prelude {
-            let code = std::fs::read_to_string(lua_code)?;
-            lua.scope(|scope| {
-                globals.raw_set("header", scope.create_any_userdata_ref_mut(&mut hv)?)?;
-                lua.load(&code).exec()
-            })?;
-        }
+        lua.scope(|scope| {
+            globals.raw_set("header", scope.create_any_userdata_ref_mut(&mut hv)?)?;
+            for path in lua_prelude {
+                let code = std::fs::read_to_string(path)?;
+                lua.load(&code).exec()?;
+            }
+            Ok(())
+        })?;
 
         let info_exps = VCFExpr::load_info_expressions(lua, &mut hv, set_expression)?;