Fix flush pass to print to and from department/account

src/move_money.rs

@@ -91,7 +91,16 @@ pub struct CsvCost {
     pub department: String,
     #[serde(serialize_with = "round_serialize")]
     pub value: f64,
-    pub pass: Option<i32>,
+}
+
+#[derive(Serialize)]
+pub struct MovedMoneyAmount {
+    pub account: String,
+    pub department: String,
+    pub value: f64,
+    pub from_account: String,
+    pub from_department: String,
+    pub pass: i32,
 }
 
 fn round_serialize<S>(x: &f64, s: S) -> Result<S::Ok, S::Error>
@@ -271,16 +280,15 @@ where
     }
 
     // Then run move_money
-    let moved = move_money_2(lines, &rules, flush_pass);
+    let moved = move_money_2(lines, &rules, flush_pass, output);
 
-    // Ouput the list moved moneys
+    if !flush_pass {
-    for money in moved {
+        // Ouput the final moneys
-        for (unit, value) in money.totals {
+        for (unit, value) in moved {
             output.serialize(CsvCost {
                 account: unit.account,
                 department: unit.department,
                 value,
-                pass: if flush_pass { Some(money.pass) } else { None },
             })?;
         }
     }
@@ -323,12 +331,6 @@ fn extract_range(from: String, to: String, options: &Vec<String>) -> HashSet<Str
 // Advantage of this approach is it can be easily extended to run on the gpu.
 pub fn move_money_1() {}
 
-pub struct MoveMoneyResult {
-    pass: i32,
-    // TODO: We want the from account and the to account
-    totals: HashMap<Unit, f64>,
-}
-
 // Approach 2:
 // Traditinoal/naive, total for each department is stored in an initial map (department -> total amount).
 // Another map is built up for each rule, and each rule is processed based on the amount in the current total
@@ -340,41 +342,39 @@ pub struct MoveMoneyResult {
 // to every. It's also much more memory efficient than approach 1.
 // TODO: Time both approaches to seee which is faster depending on the size of the input data/number of rules
 // Verdict: This is already pretty fast, at least much faster than ppm for BigDataset
-pub fn move_money_2(
+pub fn move_money_2<W>(
     initial_totals: HashMap<Unit, f64>,
     rules: &Vec<MovementRule>,
     flush_pass: bool,
-) -> Vec<MoveMoneyResult> {
+    output: &mut csv::Writer<W>,
+) -> HashMap<Unit, f64>
+where
+    W: std::io::Write,
+{
     // Note: It's potentially a bit more intensive to use cloned totals (rather than just update temp_total per rule),
     // but it's much simpler code and, and since we're only working line-by-line, it isn't really that much memory in practice
     // TODO: This is initial totals is problematic, as we may move into a cc that wasn't in the list of lines. In which case we'd need
     // to add a new line
     let mut running_total = HashMap::from(initial_totals);
     let mut temp_total = running_total.clone();
-    let mut move_money_result: Vec<MoveMoneyResult> = vec![];
     let mut current_pass = 0;
     if flush_pass {
-        move_money_result.push(MoveMoneyResult {
+        for (unit, value) in running_total.iter() {
+            output
+                .serialize(MovedMoneyAmount {
+                    account: unit.account.clone(),
+                    department: unit.department.clone(),
+                    value: *value,
+                    from_account: unit.account.clone(),
+                    from_department: unit.department.clone(),
                     pass: current_pass,
-            totals: running_total.clone(),
                 })
+                .expect("Failed to write moved amount")
+        }
     }
     for rule in rules {
         if rule.is_separator {
-            if flush_pass {
             current_pass += 1;
-                // Flush the totals at the end of this pass (more specifically the change)
-                move_money_result.push(MoveMoneyResult {
-                    pass: current_pass,
-                    totals: temp_total
-                        .iter()
-                        .map(|(unit, value)| {
-                            (unit.clone(), value - running_total.get(unit).unwrap_or(&0.))
-                        })
-                        .filter(|(_, value)| *value != 0.)
-                        .collect(),
-                });
-            }
             running_total = temp_total.clone();
         } else {
             for (unit, amount) in &running_total {
@@ -399,6 +399,18 @@ pub fn move_money_2(
                     } else {
                         unit.account.clone()
                     };
+                    if flush_pass {
+                        output
+                            .serialize(MovedMoneyAmount {
+                                department: department.clone(),
+                                account: department.clone(),
+                                value: added_amount,
+                                from_department: unit.department.clone(),
+                                from_account: unit.account.clone(),
+                                pass: current_pass,
+                            })
+                            .expect("Failed to write moved amount");
+                    }
                     *temp_total
                         .entry(Unit {
                             department,
@@ -409,12 +421,7 @@ pub fn move_money_2(
             }
         }
     }
-    move_money_result.push(MoveMoneyResult {
+    temp_total
-        pass: current_pass,
-        totals: temp_total,
-    });
-
-    move_money_result
 }
 
 #[cfg(test)]

src/overhead_allocation.rs

@@ -381,7 +381,6 @@ where
                     account: cost.account.clone(),
                     department: department.department,
                     value: department.value,
-                    pass: None,
                 })?;
             }
         }