Initial SFERA platform baseline

This commit is contained in:
2026-05-16 19:03:49 +03:00
commit 3b845c8fce
282 changed files with 55045 additions and 0 deletions
+116
View File
@@ -0,0 +1,116 @@
# This file is automatically @generated by Cargo.
# It is not intended for manual editing.
version = 4
[[package]]
name = "bsl-parser"
version = "0.1.0"
dependencies = [
"query-parser",
"serde",
"serde_json",
]
[[package]]
name = "itoa"
version = "1.0.18"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682"
[[package]]
name = "memchr"
version = "2.8.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f8ca58f447f06ed17d5fc4043ce1b10dd205e060fb3ce5b979b8ed8e59ff3f79"
[[package]]
name = "proc-macro2"
version = "1.0.106"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8fd00f0bb2e90d81d1044c2b32617f68fcb9fa3bb7640c23e9c748e53fb30934"
dependencies = [
"unicode-ident",
]
[[package]]
name = "query-parser"
version = "0.1.0"
[[package]]
name = "quote"
version = "1.0.45"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "41f2619966050689382d2b44f664f4bc593e129785a36d6ee376ddf37259b924"
dependencies = [
"proc-macro2",
]
[[package]]
name = "semantic-engine"
version = "0.1.0"
[[package]]
name = "serde"
version = "1.0.228"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9a8e94ea7f378bd32cbbd37198a4a91436180c5bb472411e48b5ec2e2124ae9e"
dependencies = [
"serde_core",
"serde_derive",
]
[[package]]
name = "serde_core"
version = "1.0.228"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "41d385c7d4ca58e59fc732af25c3983b67ac852c1a25000afe1175de458b67ad"
dependencies = [
"serde_derive",
]
[[package]]
name = "serde_derive"
version = "1.0.228"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d540f220d3187173da220f885ab66608367b6574e925011a9353e4badda91d79"
dependencies = [
"proc-macro2",
"quote",
"syn",
]
[[package]]
name = "serde_json"
version = "1.0.149"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "83fc039473c5595ace860d8c4fafa220ff474b3fc6bfdb4293327f1a37e94d86"
dependencies = [
"itoa",
"memchr",
"serde",
"serde_core",
"zmij",
]
[[package]]
name = "syn"
version = "2.0.117"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e665b8803e7b1d2a727f4023456bbbbe74da67099c585258af0ad9c5013b9b99"
dependencies = [
"proc-macro2",
"quote",
"unicode-ident",
]
[[package]]
name = "unicode-ident"
version = "1.0.24"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e6e4313cd5fcd3dad5cafa179702e2b244f760991f45397d14d4ebf38247da75"
[[package]]
name = "zmij"
version = "1.0.21"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b8848ee67ecc8aedbaf3e4122217aff892639231befc6a1b58d29fff4c2cabaa"
+7
View File
@@ -0,0 +1,7 @@
[workspace]
members = [
"crates/bsl-parser",
"crates/query-parser",
"crates/semantic-engine",
]
resolver = "2"
+9
View File
@@ -0,0 +1,9 @@
[package]
name = "bsl-parser"
version = "0.1.0"
edition = "2021"
[dependencies]
query-parser = { path = "../query-parser" }
serde = { version = "1", features = ["derive"] }
serde_json = "1"
+73
View File
@@ -0,0 +1,73 @@
use crate::keywords::{is_function_start, is_procedure_start};
use crate::models::{ParsedCall, SourceRange};
use crate::parser::{extract_name_from_header, routine_ended};
pub fn extract_calls(source: &str) -> Vec<ParsedCall> {
let mut result = Vec::new();
let mut current_routine: Option<String> = None;
for (index, line) in source.lines().enumerate() {
let line_no = index + 1;
let trimmed = strip_inline_comment(line).trim();
if is_procedure_start(trimmed) || is_function_start(trimmed) {
current_routine = Some(extract_name_from_header(trimmed));
continue;
}
if routine_ended(trimmed) {
current_routine = None;
continue;
}
let Some(caller) = current_routine.as_ref() else {
continue;
};
if let Some(callee) = simple_call_name(trimmed).or_else(|| condition_call_name(trimmed)) {
result.push(ParsedCall {
caller: caller.clone(),
callee,
source_range: SourceRange {
line_start: line_no,
line_end: line_no,
column_start: 1,
column_end: line.len() + 1,
},
});
}
}
result
}
fn simple_call_name(line: &str) -> Option<String> {
if !line.ends_with(';') || !line.contains('(') || !line.contains(')') || line.contains('.') {
return None;
}
let call_expr = line.split('=').next_back().unwrap_or(line).trim();
let name = call_expr.split('(').next()?.trim();
if name.is_empty() {
return None;
}
Some(name.to_string())
}
fn condition_call_name(line: &str) -> Option<String> {
let lowered = line.to_lowercase();
let is_condition = (lowered.starts_with("если ") && lowered.contains(" тогда"))
|| (lowered.starts_with("if ") && lowered.contains(" then"));
if !is_condition || line.contains('.') {
return None;
}
let before_then = if let Some(index) = lowered.find(" тогда") {
&line[..index]
} else if let Some(index) = lowered.find(" then") {
&line[..index]
} else {
line
};
let name_part = before_then.split('(').next()?.split_whitespace().last()?;
if name_part.eq_ignore_ascii_case("Если") || name_part.eq_ignore_ascii_case("If") {
return None;
}
Some(name_part.to_string())
}
fn strip_inline_comment(line: &str) -> &str {
line.split("//").next().unwrap_or(line)
}
+24
View File
@@ -0,0 +1,24 @@
pub fn is_procedure_start(line: &str) -> bool {
let value = line.trim_start().to_lowercase();
value.starts_with("процедура ") || value.starts_with("procedure ")
}
pub fn is_function_start(line: &str) -> bool {
let value = line.trim_start().to_lowercase();
value.starts_with("функция ") || value.starts_with("function ")
}
pub fn is_procedure_end(line: &str) -> bool {
let value = line.trim_start().to_lowercase();
value.starts_with("конецпроцедуры") || value.starts_with("endprocedure")
}
pub fn is_function_end(line: &str) -> bool {
let value = line.trim_start().to_lowercase();
value.starts_with("конецфункции") || value.starts_with("endfunction")
}
pub fn has_export(line: &str) -> bool {
let value = line.to_lowercase();
value.contains(" экспорт") || value.contains(" export")
}
+290
View File
@@ -0,0 +1,290 @@
pub mod calls;
pub mod keywords;
pub mod models;
pub mod parser;
pub mod queries;
pub mod writes;
use crate::calls::extract_calls;
use crate::models::ParsedSemanticUnit;
use crate::parser::extract_procedures;
use crate::queries::extract_queries;
use crate::writes::extract_writes;
pub fn parse_module(source_path: &str, source: &str) -> ParsedSemanticUnit {
ParsedSemanticUnit {
source_path: source_path.to_string(),
procedures: extract_procedures(source),
calls: extract_calls(source),
queries: extract_queries(source),
writes: extract_writes(source),
diagnostics: Vec::new(),
}
}
#[cfg(test)]
mod tests {
use super::parse_module;
#[test]
fn parses_procedure_calls_query_and_write() {
let source = r#"
Процедура Проведение()
ПроверитьОстатки();
Движения.ОстаткиТоваров.Записать();
КонецПроцедуры
Процедура ПроверитьОстатки()
Запрос = Новый Запрос;
Запрос.Текст =
"ВЫБРАТЬ
Остатки.Номенклатура
ИЗ
РегистрНакопления.ОстаткиТоваров КАК Остатки";
КонецПроцедуры
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.procedures.len(), 2);
assert_eq!(unit.calls.len(), 1);
assert_eq!(unit.queries.len(), 1);
assert_eq!(unit.writes.len(), 1);
assert_eq!(unit.calls[0].caller, "Проведение");
assert_eq!(unit.calls[0].callee, "ПроверитьОстатки");
assert_eq!(unit.writes[0].target, "ОстаткиТоваров");
}
#[test]
fn parses_english_function_query_call_and_write() {
let source = r#"
Procedure Posting()
CheckStock(); // inline comment
Movements.StockBalance.Write();
EndProcedure
Function CheckStock()
Query = New Query;
Query.Text =
"SELECT
Stock.Item
FROM
AccumulationRegister.StockBalance AS Stock";
EndFunction
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.procedures.len(), 2);
assert!(unit.procedures[1].is_function);
assert_eq!(unit.calls.len(), 1);
assert_eq!(unit.calls[0].callee, "CheckStock");
assert_eq!(unit.queries.len(), 1);
assert_eq!(
unit.queries[0].tables,
vec!["AccumulationRegister.StockBalance"]
);
assert!(!unit.queries[0].query_text.ends_with('"'));
assert_eq!(unit.writes.len(), 1);
assert_eq!(unit.writes[0].target, "StockBalance");
}
#[test]
fn parses_mixed_russian_and_english_keywords_in_one_module() {
let source = r#"
Procedure Posting()
ПроверитьОстатки();
Movements.StockBalance.Write();
EndProcedure
Процедура ПроверитьОстатки()
Query = New Query;
Query.Text =
"SELECT
Stock.Item
FROM
AccumulationRegister.StockBalance AS Stock";
КонецПроцедуры
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.procedures.len(), 2);
assert_eq!(unit.procedures[0].name, "Posting");
assert_eq!(unit.procedures[1].name, "ПроверитьОстатки");
assert_eq!(unit.calls.len(), 1);
assert_eq!(unit.calls[0].caller, "Posting");
assert_eq!(unit.calls[0].callee, "ПроверитьОстатки");
assert_eq!(unit.queries.len(), 1);
assert_eq!(
unit.queries[0].tables,
vec!["AccumulationRegister.StockBalance"]
);
assert_eq!(unit.writes.len(), 1);
assert_eq!(unit.writes[0].target, "StockBalance");
}
#[test]
fn parses_inline_query_assignment_with_pipe_prefixed_lines() {
let source = r#"
Процедура ПолучитьТовары()
Запрос = Новый Запрос;
Запрос.Текст = "ВЫБРАТЬ
|Товары.Ссылка
|ИЗ
|Справочник.Номенклатура КАК Товары";
КонецПроцедуры
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.queries.len(), 1);
assert_eq!(unit.queries[0].tables, vec!["Справочник.Номенклатура"]);
assert!(unit.queries[0].query_text.starts_with("ВЫБРАТЬ"));
assert!(!unit.queries[0].query_text.contains("|ИЗ"));
}
#[test]
fn parses_query_from_and_table_on_same_line() {
let source = r#"
Процедура ПолучитьКонтрагентов()
Запрос = Новый Запрос;
Запрос.Текст = "ВЫБРАТЬ
|Контрагенты.Ссылка
|ИЗ Справочник.Контрагенты КАК Контрагенты";
КонецПроцедуры
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.queries.len(), 1);
assert_eq!(unit.queries[0].tables, vec!["Справочник.Контрагенты"]);
}
#[test]
fn parses_assignment_function_calls() {
let source = r#"
Процедура Проведение()
МожноПроводить = ПроверитьОстатки();
КонецПроцедуры
Функция ПроверитьОстатки()
Возврат Истина;
КонецФункции
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.calls.len(), 1);
assert_eq!(unit.calls[0].caller, "Проведение");
assert_eq!(unit.calls[0].callee, "ПроверитьОстатки");
}
#[test]
fn parses_condition_function_calls() {
let source = r#"
Процедура Проведение()
Если ПроверитьОстатки() Тогда
Движения.ОстаткиТоваров.Записать();
КонецЕсли;
КонецПроцедуры
Функция ПроверитьОстатки()
Возврат Истина;
КонецФункции
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.calls.len(), 1);
assert_eq!(unit.calls[0].caller, "Проведение");
assert_eq!(unit.calls[0].callee, "ПроверитьОстатки");
}
#[test]
fn parses_join_tables_from_query_text() {
let source = r#"
Процедура ПолучитьЗаказы()
Запрос = Новый Запрос;
Запрос.Текст = "ВЫБРАТЬ
|Заказы.Ссылка,
|Контрагенты.Наименование
|ИЗ Документ.ЗаказПокупателя КАК Заказы
|ЛЕВОЕ СОЕДИНЕНИЕ Справочник.Контрагенты КАК Контрагенты
|ПО Заказы.Контрагент = Контрагенты.Ссылка";
КонецПроцедуры
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(
unit.queries[0].tables,
vec!["Документ.ЗаказПокупателя", "Справочник.Контрагенты"]
);
}
#[test]
fn parses_object_write_targets_from_create_assignments() {
let source = r#"
Процедура СоздатьНоменклатуру()
Элемент = Справочники.Номенклатура.СоздатьЭлемент();
Элемент.Записать();
КонецПроцедуры
Procedure CreateOrder()
Order = Documents.CustomerOrder.CreateDocument();
Order.Write();
EndProcedure
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.writes.len(), 2);
assert_eq!(unit.writes[0].target, "Справочник.Номенклатура");
assert_eq!(unit.writes[0].write_type, "OBJECT_WRITE");
assert_eq!(unit.writes[1].target, "Документ.CustomerOrder");
assert_eq!(unit.writes[1].write_type, "OBJECT_WRITE");
}
#[test]
fn parses_recordset_write_targets_from_create_assignments() {
let source = r#"
Процедура ЗаписатьЦены()
Набор = РегистрыСведений.Цены.СоздатьНаборЗаписей();
Набор.Записать();
КонецПроцедуры
Procedure WriteBalances()
Records = AccumulationRegisters.StockBalance.CreateRecordSet();
Records.Write();
EndProcedure
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.writes.len(), 2);
assert_eq!(unit.writes[0].target, "РегистрСведений.Цены");
assert_eq!(unit.writes[0].write_type, "REGISTER_WRITE");
assert_eq!(unit.writes[1].target, "РегистрНакопления.StockBalance");
assert_eq!(unit.writes[1].write_type, "REGISTER_WRITE");
}
#[test]
fn parses_calls_and_writes_inside_control_flow_blocks() {
let source = r#"
Процедура Проведение()
Для Каждого Строка Из Товары Цикл
ПроверитьСтроку(Строка);
КонецЦикла;
Попытка
Движения.ОстаткиТоваров.Записать();
Исключение
СообщитьОбОшибке();
КонецПопытки;
КонецПроцедуры
Процедура ПроверитьСтроку(Строка)
КонецПроцедуры
Процедура СообщитьОбОшибке()
КонецПроцедуры
"#;
let unit = parse_module("module.bsl", source);
assert_eq!(unit.calls.len(), 2);
assert_eq!(unit.calls[0].callee, "ПроверитьСтроку");
assert_eq!(unit.calls[1].callee, "СообщитьОбОшибке");
assert_eq!(unit.writes.len(), 1);
assert_eq!(unit.writes[0].target, "ОстаткиТоваров");
}
}
+33
View File
@@ -0,0 +1,33 @@
use std::env;
use std::fs;
use std::io::{self, Read};
use bsl_parser::parse_module;
fn main() {
let args: Vec<String> = env::args().collect();
let source_path = args.get(1).map(String::as_str).unwrap_or("-");
let source = match args.get(1) {
Some(path) => fs::read_to_string(path).unwrap_or_else(|error| {
eprintln!("failed to read {path}: {error}");
std::process::exit(2);
}),
None => {
let mut buffer = String::new();
io::stdin()
.read_to_string(&mut buffer)
.unwrap_or_else(|error| {
eprintln!("failed to read stdin: {error}");
std::process::exit(2);
});
buffer
}
};
let unit = parse_module(source_path, &source);
serde_json::to_writer_pretty(io::stdout(), &unit).unwrap_or_else(|error| {
eprintln!("failed to serialize parse result: {error}");
std::process::exit(2);
});
println!();
}
+59
View File
@@ -0,0 +1,59 @@
use serde::Serialize;
#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
pub struct SourceRange {
pub line_start: usize,
pub line_end: usize,
pub column_start: usize,
pub column_end: usize,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
pub struct ParsedProcedure {
pub name: String,
pub export: bool,
pub is_function: bool,
pub parameters: Vec<String>,
pub source_range: SourceRange,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
pub struct ParsedCall {
pub caller: String,
pub callee: String,
pub source_range: SourceRange,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
pub struct ParsedQuery {
pub owner_procedure: String,
pub query_text: String,
pub tables: Vec<String>,
pub source_range: SourceRange,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
pub struct ParsedWrite {
pub owner_procedure: String,
pub target: String,
pub write_type: String,
pub source_range: SourceRange,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
pub struct ParseDiagnostic {
pub code: String,
pub message: String,
pub severity: String,
pub source_range: Option<SourceRange>,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
pub struct ParsedSemanticUnit {
pub source_path: String,
pub procedures: Vec<ParsedProcedure>,
pub calls: Vec<ParsedCall>,
pub queries: Vec<ParsedQuery>,
pub writes: Vec<ParsedWrite>,
pub diagnostics: Vec<ParseDiagnostic>,
}
+85
View File
@@ -0,0 +1,85 @@
use crate::keywords::{
has_export, is_function_end, is_function_start, is_procedure_end, is_procedure_start,
};
use crate::models::{ParsedProcedure, SourceRange};
pub fn extract_procedures(source: &str) -> Vec<ParsedProcedure> {
let mut result = Vec::new();
let mut current_index: Option<usize> = None;
for (index, line) in source.lines().enumerate() {
let line_no = index + 1;
if is_procedure_start(line) || is_function_start(line) {
let is_function = is_function_start(line);
let name = extract_routine_name(line, is_function);
result.push(ParsedProcedure {
name,
export: has_export(line),
is_function,
parameters: extract_parameters(line),
source_range: SourceRange {
line_start: line_no,
line_end: line_no,
column_start: 1,
column_end: line.len() + 1,
},
});
current_index = Some(result.len() - 1);
} else if routine_ended(line) {
if let Some(procedure_index) = current_index.take() {
result[procedure_index].source_range.line_end = line_no;
result[procedure_index].source_range.column_end = line.len() + 1;
}
}
}
result
}
pub fn routine_ended(line: &str) -> bool {
is_procedure_end(line) || is_function_end(line)
}
pub fn extract_name_from_header(line: &str) -> String {
line.split_whitespace()
.nth(1)
.unwrap_or("")
.split('(')
.next()
.unwrap_or("")
.to_string()
}
fn extract_routine_name(line: &str, is_function: bool) -> String {
let trimmed = line.trim();
let prefixes = if is_function {
vec!["Функция", "Function"]
} else {
vec!["Процедура", "Procedure"]
};
for prefix in prefixes {
if trimmed.to_lowercase().starts_with(&prefix.to_lowercase()) {
return trimmed[prefix.len()..]
.trim()
.split('(')
.next()
.unwrap_or("")
.trim()
.to_string();
}
}
String::new()
}
fn extract_parameters(line: &str) -> Vec<String> {
let Some(start) = line.find('(') else {
return Vec::new();
};
let Some(end) = line[start + 1..].find(')') else {
return Vec::new();
};
let params = &line[start + 1..start + 1 + end];
params
.split(',')
.map(|p| p.trim().to_string())
.filter(|p| !p.is_empty())
.collect()
}
+121
View File
@@ -0,0 +1,121 @@
use crate::keywords::{is_function_start, is_procedure_start};
use crate::models::{ParsedQuery, SourceRange};
use crate::parser::{extract_name_from_header, routine_ended};
use query_parser::extract_tables;
pub fn extract_queries(source: &str) -> Vec<ParsedQuery> {
let mut result = Vec::new();
let mut current_routine: Option<String> = None;
let mut collecting = false;
let mut query_lines: Vec<String> = Vec::new();
let mut query_start_line = 0;
for (index, line) in source.lines().enumerate() {
let line_no = index + 1;
let trimmed = strip_inline_comment(line).trim();
if is_procedure_start(trimmed) || is_function_start(trimmed) {
current_routine = Some(extract_name_from_header(trimmed));
continue;
}
if routine_ended(trimmed) {
current_routine = None;
collecting = false;
query_lines.clear();
continue;
}
if trimmed.contains(".Текст") || trimmed.to_lowercase().contains(".text") {
collecting = true;
query_start_line = line_no;
if let Some(inline_query) = query_text_after_assignment(trimmed) {
query_lines.push(clean_query_line(inline_query));
if trimmed.ends_with(';') {
push_query(
&mut result,
current_routine.as_ref(),
&query_lines,
query_start_line,
line_no,
line.len() + 1,
);
collecting = false;
query_lines.clear();
}
}
continue;
}
if collecting {
query_lines.push(clean_query_line(trimmed));
if trimmed.ends_with(';') {
push_query(
&mut result,
current_routine.as_ref(),
&query_lines,
query_start_line,
line_no,
line.len() + 1,
);
collecting = false;
query_lines.clear();
}
}
}
result
}
fn push_query(
result: &mut Vec<ParsedQuery>,
owner: Option<&String>,
query_lines: &[String],
line_start: usize,
line_end: usize,
column_end: usize,
) {
let Some(owner) = owner else {
return;
};
let query_text = query_lines
.iter()
.filter(|line| !line.is_empty())
.cloned()
.collect::<Vec<_>>()
.join("\n");
let tables = extract_tables(&query_text);
result.push(ParsedQuery {
owner_procedure: owner.clone(),
query_text,
tables,
source_range: SourceRange {
line_start,
line_end,
column_start: 1,
column_end,
},
});
}
fn clean_query_line(line: &str) -> String {
let mut value = line
.trim()
.trim_end_matches(';')
.trim()
.trim_matches('"')
.trim();
if let Some(stripped) = value.strip_prefix('|') {
value = stripped.trim();
}
value.trim_matches('"').trim().to_string()
}
fn query_text_after_assignment(line: &str) -> Option<&str> {
let (_, value) = line.split_once('=')?;
let trimmed = value.trim();
if trimmed.is_empty() {
None
} else {
Some(trimmed)
}
}
fn strip_inline_comment(line: &str) -> &str {
line.split("//").next().unwrap_or(line)
}
+159
View File
@@ -0,0 +1,159 @@
use crate::keywords::{is_function_start, is_procedure_start};
use crate::models::{ParsedWrite, SourceRange};
use crate::parser::{extract_name_from_header, routine_ended};
use std::collections::HashMap;
pub fn extract_writes(source: &str) -> Vec<ParsedWrite> {
let mut result = Vec::new();
let mut current_routine: Option<String> = None;
let mut object_targets: HashMap<String, String> = HashMap::new();
for (index, line) in source.lines().enumerate() {
let line_no = index + 1;
let trimmed = strip_inline_comment(line).trim();
if is_procedure_start(trimmed) || is_function_start(trimmed) {
current_routine = Some(extract_name_from_header(trimmed));
object_targets.clear();
continue;
}
if routine_ended(trimmed) {
current_routine = None;
continue;
}
let Some(owner) = current_routine.as_ref() else {
continue;
};
let lowered = trimmed.to_lowercase();
if let Some((variable, target)) = extract_object_create_target(trimmed) {
object_targets.insert(variable.to_lowercase(), target);
}
if (trimmed.contains("Движения.") && trimmed.contains(".Записать"))
|| (lowered.contains("movements.") && lowered.contains(".write"))
{
result.push(ParsedWrite {
owner_procedure: owner.clone(),
target: extract_register_target(trimmed),
write_type: "REGISTER_WRITE".to_string(),
source_range: SourceRange {
line_start: line_no,
line_end: line_no,
column_start: 1,
column_end: line.len() + 1,
},
});
} else if let Some((target, write_type)) =
extract_object_write_target(trimmed, &object_targets)
{
result.push(ParsedWrite {
owner_procedure: owner.clone(),
target,
write_type,
source_range: SourceRange {
line_start: line_no,
line_end: line_no,
column_start: 1,
column_end: line.len() + 1,
},
});
} else if trimmed.contains(".Записать()") || trimmed.contains(".Write()") {
result.push(ParsedWrite {
owner_procedure: owner.clone(),
target: "unknown".to_string(),
write_type: "OBJECT_WRITE".to_string(),
source_range: SourceRange {
line_start: line_no,
line_end: line_no,
column_start: 1,
column_end: line.len() + 1,
},
});
}
}
result
}
fn extract_register_target(line: &str) -> String {
let lowered = line.to_lowercase();
let after = if let Some(value) = line.split("Движения.").nth(1) {
value
} else if let Some(index) = lowered.find("movements.") {
&line[index + "movements.".len()..]
} else {
return "unknown".to_string();
};
after.split('.').next().unwrap_or("unknown").to_string()
}
fn extract_object_create_target(line: &str) -> Option<(String, String)> {
let (variable, expression) = line.split_once('=')?;
let variable = variable.trim();
let expression = expression.trim();
let parts: Vec<&str> = expression.split('.').collect();
if parts.len() < 3 {
return None;
}
let prefix = if parts[0].eq_ignore_ascii_case("Справочники")
|| parts[0].eq_ignore_ascii_case("Catalogs")
{
"Справочник"
} else if parts[0].eq_ignore_ascii_case("Документы")
|| parts[0].eq_ignore_ascii_case("Documents")
{
"Документ"
} else if parts[0].eq_ignore_ascii_case("РегистрыСведений")
|| parts[0].eq_ignore_ascii_case("InformationRegisters")
{
"РегистрСведений"
} else if parts[0].eq_ignore_ascii_case("РегистрыНакопления")
|| parts[0].eq_ignore_ascii_case("AccumulationRegisters")
{
"РегистрНакопления"
} else if parts[0].eq_ignore_ascii_case("РегистрыБухгалтерии")
|| parts[0].eq_ignore_ascii_case("AccountingRegisters")
{
"РегистрБухгалтерии"
} else if parts[0].eq_ignore_ascii_case("РегистрыРасчета")
|| parts[0].eq_ignore_ascii_case("CalculationRegisters")
{
"РегистрРасчета"
} else {
return None;
};
let method = parts[2].split('(').next()?.trim();
let valid_method = method.eq_ignore_ascii_case("СоздатьЭлемент")
|| method.eq_ignore_ascii_case("CreateItem")
|| method.eq_ignore_ascii_case("СоздатьДокумент")
|| method.eq_ignore_ascii_case("CreateDocument")
|| method.eq_ignore_ascii_case("СоздатьНаборЗаписей")
|| method.eq_ignore_ascii_case("CreateRecordSet");
if !valid_method || variable.is_empty() || parts[1].trim().is_empty() {
return None;
}
Some((
variable.to_string(),
format!("{}.{}", prefix, parts[1].trim()),
))
}
fn extract_object_write_target(
line: &str,
object_targets: &HashMap<String, String>,
) -> Option<(String, String)> {
let (variable, method) = line.split_once('.')?;
let method_name = method.split('(').next()?.trim();
let is_write =
method_name.eq_ignore_ascii_case("Записать") || method_name.eq_ignore_ascii_case("Write");
if !is_write {
return None;
}
let target = object_targets.get(&variable.trim().to_lowercase())?.clone();
let write_type = if target.to_lowercase().starts_with("регистр") {
"REGISTER_WRITE"
} else {
"OBJECT_WRITE"
};
Some((target, write_type.to_string()))
}
fn strip_inline_comment(line: &str) -> &str {
line.split("//").next().unwrap_or(line)
}
+4
View File
@@ -0,0 +1,4 @@
[package]
name = "query-parser"
version = "0.1.0"
edition = "2021"
+143
View File
@@ -0,0 +1,143 @@
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct ParsedQueryText {
pub tables: Vec<String>,
}
pub fn parse_query_text(query_text: &str) -> ParsedQueryText {
ParsedQueryText {
tables: extract_tables(query_text),
}
}
pub fn extract_tables(query_text: &str) -> Vec<String> {
let mut tables = Vec::new();
let lines: Vec<String> = query_text.lines().map(clean_query_line).collect();
for (index, line) in lines.iter().enumerate() {
if let Some(table) = table_after_from(line) {
push_unique(&mut tables, table);
} else if is_standalone_from(line) {
if let Some(next_line) = lines.get(index + 1) {
if let Some(table) = first_table_token(next_line) {
push_unique(&mut tables, table);
}
}
}
if let Some(table) = table_after_join(line) {
push_unique(&mut tables, table);
}
}
tables
}
fn clean_query_line(line: &str) -> String {
let mut value = line
.trim()
.trim_end_matches(';')
.trim()
.trim_matches('"')
.trim();
if let Some(stripped) = value.strip_prefix('|') {
value = stripped.trim();
}
value.trim_matches('"').trim().to_string()
}
fn is_standalone_from(line: &str) -> bool {
line.eq_ignore_ascii_case("ИЗ") || line.eq_ignore_ascii_case("FROM")
}
fn table_after_from(line: &str) -> Option<String> {
let trimmed = line.trim();
let upper = trimmed.to_uppercase();
let rest = if upper.starts_with("ИЗ ") {
&trimmed["ИЗ".len()..]
} else if upper.starts_with("FROM ") {
&trimmed["FROM".len()..]
} else {
return None;
};
first_table_token(rest)
}
fn table_after_join(line: &str) -> Option<String> {
let normalized = line.replace('\t', " ");
let parts: Vec<&str> = normalized.split_whitespace().collect();
let index = parts.iter().position(|part| {
part.eq_ignore_ascii_case("JOIN") || part.eq_ignore_ascii_case("СОЕДИНЕНИЕ")
})?;
let table = parts.get(index + 1)?.trim_matches(',');
if table.is_empty() {
None
} else {
Some(table.to_string())
}
}
fn first_table_token(line: &str) -> Option<String> {
let table = line.trim().split_whitespace().next()?.trim_matches(',');
if table.is_empty() {
None
} else {
Some(table.to_string())
}
}
fn push_unique(values: &mut Vec<String>, value: String) {
if !values.iter().any(|existing| existing == &value) {
values.push(value);
}
}
#[cfg(test)]
mod tests {
use super::{extract_tables, parse_query_text};
#[test]
fn extracts_table_after_standalone_from() {
let tables = extract_tables(
r#"
ВЫБРАТЬ
Остатки.Номенклатура
ИЗ
РегистрНакопления.ОстаткиТоваров КАК Остатки
"#,
);
assert_eq!(tables, vec!["РегистрНакопления.ОстаткиТоваров"]);
}
#[test]
fn extracts_inline_from_and_join_tables() {
let query = parse_query_text(
r#"
SELECT
Orders.Ref,
Customers.Description
FROM Document.CustomerOrder AS Orders
LEFT JOIN Catalog.Customers AS Customers
ON Orders.Customer = Customers.Ref
"#,
);
assert_eq!(
query.tables,
vec!["Document.CustomerOrder", "Catalog.Customers"]
);
}
#[test]
fn cleans_pipe_prefixed_1c_query_lines() {
let tables = extract_tables(
r#"
|ВЫБРАТЬ
|Контрагенты.Ссылка
|ИЗ Справочник.Контрагенты КАК Контрагенты
"#,
);
assert_eq!(tables, vec!["Справочник.Контрагенты"]);
}
}
+4
View File
@@ -0,0 +1,4 @@
[package]
name = "semantic-engine"
version = "0.1.0"
edition = "2021"
+181
View File
@@ -0,0 +1,181 @@
use std::collections::{BTreeMap, BTreeSet};
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct SemanticNode {
pub lineage_id: String,
pub kind: String,
pub name: String,
pub qualified_name: String,
}
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct SemanticEdge {
pub edge_id: String,
pub kind: String,
pub source_lineage: String,
pub target_lineage: String,
}
#[derive(Debug, Default, Clone, PartialEq, Eq)]
pub struct SemanticGraph {
nodes: BTreeMap<String, SemanticNode>,
outgoing: BTreeMap<String, Vec<SemanticEdge>>,
incoming: BTreeMap<String, Vec<SemanticEdge>>,
}
impl SemanticGraph {
pub fn new(nodes: Vec<SemanticNode>, edges: Vec<SemanticEdge>) -> Self {
let nodes_by_lineage = nodes
.into_iter()
.map(|node| (node.lineage_id.clone(), node))
.collect();
let mut graph = Self {
nodes: nodes_by_lineage,
outgoing: BTreeMap::new(),
incoming: BTreeMap::new(),
};
for edge in edges {
graph.add_edge(edge);
}
graph
}
pub fn add_edge(&mut self, edge: SemanticEdge) {
self.outgoing
.entry(edge.source_lineage.clone())
.or_default()
.push(edge.clone());
self.incoming
.entry(edge.target_lineage.clone())
.or_default()
.push(edge);
}
pub fn find_callers(&self, routine_name: &str) -> Vec<&SemanticNode> {
let target_lineages = self.routine_lineages(routine_name);
let caller_lineages = target_lineages
.iter()
.flat_map(|lineage| self.incoming.get(lineage).into_iter().flatten())
.filter(|edge| edge.kind == "CALLS")
.map(|edge| edge.source_lineage.as_str())
.collect::<BTreeSet<_>>();
caller_lineages
.iter()
.filter_map(|lineage| self.nodes.get(*lineage))
.collect()
}
pub fn find_callees(&self, routine_name: &str) -> Vec<&SemanticNode> {
self.routine_lineages(routine_name)
.iter()
.flat_map(|lineage| self.outgoing.get(lineage).into_iter().flatten())
.filter(|edge| edge.kind == "CALLS")
.filter_map(|edge| self.nodes.get(&edge.target_lineage))
.collect()
}
pub fn find_writes(&self, routine_name: &str) -> Vec<&SemanticNode> {
self.targets_by_edge_kind(routine_name, "WRITES")
}
pub fn find_reads(&self, routine_name: &str) -> Vec<&SemanticNode> {
let query_lineages = self
.routine_lineages(routine_name)
.iter()
.flat_map(|lineage| self.outgoing.get(lineage).into_iter().flatten())
.filter(|edge| edge.kind == "OWNS_QUERY")
.map(|edge| edge.target_lineage.as_str())
.collect::<BTreeSet<_>>();
query_lineages
.iter()
.flat_map(|lineage| self.outgoing.get(*lineage).into_iter().flatten())
.filter(|edge| edge.kind == "READS_TABLE")
.filter_map(|edge| self.nodes.get(&edge.target_lineage))
.collect()
}
fn targets_by_edge_kind(&self, routine_name: &str, kind: &str) -> Vec<&SemanticNode> {
self.routine_lineages(routine_name)
.iter()
.flat_map(|lineage| self.outgoing.get(lineage).into_iter().flatten())
.filter(|edge| edge.kind == kind)
.filter_map(|edge| self.nodes.get(&edge.target_lineage))
.collect()
}
fn routine_lineages(&self, routine_name: &str) -> BTreeSet<String> {
let wanted = routine_name.to_lowercase();
self.nodes
.values()
.filter(|node| {
matches!(node.kind.as_str(), "PROCEDURE" | "FUNCTION")
&& node.name.to_lowercase() == wanted
})
.map(|node| node.lineage_id.clone())
.collect()
}
}
#[cfg(test)]
mod tests {
use super::{SemanticEdge, SemanticGraph, SemanticNode};
#[test]
fn resolves_callers_callees_reads_and_writes() {
let graph = SemanticGraph::new(
vec![
node("routine.post", "PROCEDURE", "Проведение"),
node("routine.check", "FUNCTION", "ПроверитьОстатки"),
node("query.check.1", "QUERY", "ПроверитьОстатки.query1"),
node("table.stock", "REGISTER", "ОстаткиТоваров"),
],
vec![
edge("e1", "CALLS", "routine.post", "routine.check"),
edge("e2", "WRITES", "routine.post", "table.stock"),
edge("e3", "OWNS_QUERY", "routine.check", "query.check.1"),
edge("e4", "READS_TABLE", "query.check.1", "table.stock"),
],
);
assert_eq!(
names(graph.find_callees("Проведение")),
vec!["ПроверитьОстатки"]
);
assert_eq!(
names(graph.find_callers("ПроверитьОстатки")),
vec!["Проведение"]
);
assert_eq!(
names(graph.find_writes("Проведение")),
vec!["ОстаткиТоваров"]
);
assert_eq!(
names(graph.find_reads("ПроверитьОстатки")),
vec!["ОстаткиТоваров"]
);
}
fn node(lineage_id: &str, kind: &str, name: &str) -> SemanticNode {
SemanticNode {
lineage_id: lineage_id.to_string(),
kind: kind.to_string(),
name: name.to_string(),
qualified_name: name.to_string(),
}
}
fn edge(edge_id: &str, kind: &str, source: &str, target: &str) -> SemanticEdge {
SemanticEdge {
edge_id: edge_id.to_string(),
kind: kind.to_string(),
source_lineage: source.to_string(),
target_lineage: target.to_string(),
}
}
fn names(nodes: Vec<&SemanticNode>) -> Vec<String> {
nodes.into_iter().map(|node| node.name.clone()).collect()
}
}