john
/
pdf-parsers
1
0
Fork 0
Code

Add Caruna/Helen composite parser

This commit is contained in:
John Ahlroos 2021-04-20 14:22:33 +03:00
parent 9e883631ec
commit 8cd1376240
Signed by: john
GPG Key ID: 258D0F70DB84CD5D
12 changed files with 135 additions and 94 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.gitignore vendored
View File

@ -1,2 +1,3 @@
.gradle
.idea
build

11
README.md
View File

@ -7,11 +7,18 @@ PDF parser to parse Caruna invoices.
Usage: ``./gradlew :caruna-invoice:run --args="/path/to/pdf"``
### Helen Invoice Parser
PDF parser to parse Helen invoices.
Usage: ``./gradlew :helen-invoice:run --args="/path/to/pdf"``
Usage: ``./gradlew :helen-invoice:run --args="/path/to/pdf <month>:dayKwh <month>:nightKwh"``
### Helen/Caruna Composite Parser
PDF parser to combine Helen/Caruna invoices into single lines
Usage: ``./gradlew gradle :composite-parsers:run --args="<caruna invoice>.pdf <helen-invoice>.pdf""``

BIN
caruna-invoice/build/classes/java/main/com/devsoap/parsers/caruna/Parser.class
View File

Binary file not shown.

2
caruna-invoice/build/tmp/compileJava/source-classes-mapping.txt
View File

@ -1,2 +0,0 @@
com/devsoap/parsers/caruna/Parser.java
com.devsoap.parsers.caruna.Parser

81
caruna-invoice/src/main/java/com/devsoap/parsers/helen/Parser.java → caruna-invoice/src/main/java/com/devsoap/parsers/caruna/Parser.java
View File

@ -1,19 +1,17 @@
package com.devsoap.parsers.helen;
package com.devsoap.parsers.caruna;
import com.itextpdf.kernel.pdf.PdfDocument;
import com.itextpdf.kernel.pdf.PdfReader;
import com.itextpdf.kernel.pdf.canvas.parser.PdfTextExtractor;
import java.io.FileNotFoundException;
import java.io.IOException;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.time.LocalDate;
import java.time.format.DateTimeFormatter;
import java.time.format.FormatStyle;
import java.time.format.TextStyle;
import java.util.Locale;
import java.util.Objects;
import java.util.Scanner;
import java.util.*;
import java.util.regex.Pattern;
public class Parser {
@ -29,78 +27,81 @@ public class Parser {
.ofLocalizedDate(FormatStyle.SHORT)
.withLocale(FI_LOCALE);
public static class Period {
public double basicPay = 0.0;
public double transferDayPrice = 0.0;
public double transferDayTotal = 0.0;
public int transferDayKwh = 0;
public double transferNightPrice = 0.0;
public double transferNightTotal = 0.0;
public int transferNightKwh = 0;
public double tax = 0.0;
}
public static void main(String[] args) {
var filename = args[0];
var file = Paths.get(filename);
System.out.println("Kuukausi,Perusmaksu (energia),Perusmaksu (siirto),Päiväenergia (kWh),Päiväenergia " +
"(EUR),Yöenergia (kWh),Yöenergia (EUR),Päiväsiirto (kWh),Päiväsiirto (EUR),Yösiirto (kWh)" +
",Yösiirto (EUR),Vero");
parse(file).forEach((n,p) -> {
var csv = String.format("%s,,%.02f,,,,,%d, %.02f,%d, %.02f, %.02f",
n, p.basicPay, p.transferDayKwh, p.transferDayTotal, p.transferNightKwh, p.transferNightTotal, p.tax);
System.out.println(csv);
});
}
public static Map<String, Period> parse(Path file) {
try(var reader = new PdfReader(file.toFile())) {
var document = new PdfDocument(reader);
var page2 = document.getPage(2);
var text = PdfTextExtractor.getTextFromPage(page2);
var scanner = new Scanner(text);
var month = "";
var basicPay = 0.0;
var transferDayPrice = 0.0;
var transferDayTotal = 0.0;
var transforDayKwh = 0L;
var transferNightPrice = 0.0;
var transferNightTotal = 0.0;
var transforNightKwh = 0L;
var tax = 0.0;
System.out.println("Kuukausi,Perusmaksu (energia),Perusmaksu (siirto),Päiväenergia (kWh),Päiväenergia " +
"(EUR),Yöenergia (kWh),Yöenergia (EUR),Päiväsiirto (kWh),Päiväsiirto (EUR),Yösiirto (kWh)" +
",Yösiirto (EUR),Vero");
var periods = new HashMap<String, Period>();
Period currentPeriod = null;
while(scanner.hasNextLine()) {
var line = scanner.nextLine();
if(DATE_RANGE_PATTERN.asPredicate().test(line)) {
if(!Objects.equals(month, "")) {
var csv = String.format("%s,,%.02f,,,,,%d, %.02f,%d, %.02f, %.02f",
month, basicPay, transforDayKwh, transferDayTotal, transforNightKwh, transferNightTotal, tax);
System.out.println(csv);
}
var matcher = DATE_RANGE_PATTERN.matcher(line);
while(matcher.find()) {
month = LocalDate.from( FI_DATE.parse(matcher.group(1)))
var month = LocalDate.from( FI_DATE.parse(matcher.group(1)))
.getMonth()
.getDisplayName(TextStyle.FULL, new Locale("FI","fi"));
month = month.substring(0,1).toUpperCase() + month.substring(1, month.length()-2);
currentPeriod = periods.computeIfAbsent(month, s -> new Period());
}
} else if(PERUSMAKSU_PATTERN.asPredicate().test(line)) {
var matcher = PERUSMAKSU_PATTERN.matcher(line);
while (matcher.find()) {
basicPay = Double.parseDouble(matcher.group(1).replace(",", "."));
currentPeriod.basicPay = Double.parseDouble(matcher.group(1).replace(",", "."));
}
} else if(P_SIIRTO_PATTERN.asPredicate().test(line)) {
var matcher = P_SIIRTO_PATTERN.matcher(line);
while (matcher.find()) {
transferDayPrice = Double.parseDouble(matcher.group(1).replace(",", ".")) / 100.0;
transferDayTotal = Double.parseDouble(matcher.group(2).replace(",", "."));
transforDayKwh = Math.round(transferDayTotal / transferDayPrice);
currentPeriod.transferDayPrice = Double.parseDouble(matcher.group(1).replace(",", ".")) / 100.0;
currentPeriod.transferDayTotal = Double.parseDouble(matcher.group(2).replace(",", "."));
currentPeriod.transferDayKwh = (int) Math.round(currentPeriod.transferDayTotal / currentPeriod.transferDayPrice);
}
} else if(O_SIIRTO_PATTERN.asPredicate().test(line)) {
var matcher = O_SIIRTO_PATTERN.matcher(line);
while (matcher.find()) {
transferNightPrice = Double.parseDouble(matcher.group(1).replace(",", ".")) / 100.0;
transferNightTotal = Double.parseDouble(matcher.group(2).replace(",", "."));
transforNightKwh = Math.round(transferNightTotal / transferNightPrice);
currentPeriod.transferNightPrice = Double.parseDouble(matcher.group(1).replace(",", ".")) / 100.0;
currentPeriod.transferNightTotal = Double.parseDouble(matcher.group(2).replace(",", "."));
currentPeriod.transferNightKwh = (int) Math.round(currentPeriod.transferNightTotal / currentPeriod.transferNightPrice);
}
} else if(TAX_PATTERN.asPredicate().test(line)) {
var matcher = TAX_PATTERN.matcher(line);
while (matcher.find()) {
tax = Double.parseDouble(matcher.group(1).replace(",", "."));
currentPeriod.tax = Double.parseDouble(matcher.group(1).replace(",", "."));
}
}
}
var csv = String.format("%s,,%.02f,,,,,%d, %.02f,%d, %.02f, %.02f",
month, basicPay, transforDayKwh, transferDayTotal, transforNightKwh, transferNightTotal, tax);
System.out.println(csv);
} catch (FileNotFoundException e) {
e.printStackTrace();
return periods;
} catch (IOException e) {
e.printStackTrace();
throw new RuntimeException(e);
}
}
}

6
composite-parsers/build.gradle Normal file
View File

@ -0,0 +1,6 @@
dependencies {
implementation project(':caruna-invoice')
implementation project(':helen-invoice')
}
application.mainClass='com.devsoap.parsers.composites.CarunaHelenParser'

40
composite-parsers/src/main/java/com/devsoap/parsers/composites/CarunaHelenParser.java Normal file
View File

@ -0,0 +1,40 @@
package com.devsoap.parsers.composites;
import java.nio.file.Path;
import java.util.ArrayList;
import java.util.HashSet;
import java.util.Map;
import java.util.stream.Collectors;
public class CarunaHelenParser {
public static void main(String[] args) {
var carunaFile = Path.of("/home/john/Downloads/caruna-lasku.pdf");
var helenFile = Path.of("/home/john/Downloads/helen-lasku.pdf");
var carunaPeriods = com.devsoap.parsers.caruna.Parser.parse(carunaFile);
var nightSiirto = carunaPeriods.entrySet().stream()
.collect(Collectors.toMap(Map.Entry::getKey, e -> e.getValue().transferNightKwh));
var daySiirto = carunaPeriods.entrySet().stream()
.collect(Collectors.toMap(Map.Entry::getKey, e -> e.getValue().transferDayKwh));
var helenPeriods = com.devsoap.parsers.helen.Parser.parse(helenFile, daySiirto, nightSiirto);
System.out.println("Kuukausi,Perusmaksu (energia),Perusmaksu (siirto),Päiväenergia (kWh),Päiväenergia " +
"(EUR),Yöenergia (kWh),Yöenergia (EUR),Päiväsiirto (kWh),Päiväsiirto (EUR),Yösiirto (kWh)" +
",Yösiirto (EUR),Vero");
var months = new HashSet<>(carunaPeriods.keySet());
months.addAll(helenPeriods.keySet());
months.forEach(month -> {
var hp = helenPeriods.getOrDefault(month, new com.devsoap.parsers.helen.Parser.Period());
var cp = carunaPeriods.getOrDefault(month, new com.devsoap.parsers.caruna.Parser.Period());
var csv = String.format("%s,%.02f,%.02f,%d,%.02f,%d,%.02f,%d,%.02f,%d,%.02f,%.02f",
month, hp.basicPay, cp.basicPay, hp.dayEnergy, hp.dayEnergyEur, hp.nightEnergy,
hp.nightEnergyEur, cp.transferDayKwh, cp.transferDayTotal, cp.transferNightKwh,
cp.transferNightTotal, cp.tax);
csv = csv.replace(",0,",",,").replace(",0.00",",");
System.out.println(csv);
});
}
}

BIN
helen-invoice/build/classes/java/main/com/devsoap/parsers/helen/Parser$Period.class
View File

Binary file not shown.

BIN
helen-invoice/build/classes/java/main/com/devsoap/parsers/helen/Parser.class
View File

Binary file not shown.

3
helen-invoice/build/tmp/compileJava/source-classes-mapping.txt
View File

@ -1,3 +0,0 @@
com/devsoap/parsers/helen/Parser.java
com.devsoap.parsers.helen.Parser
com.devsoap.parsers.helen.Parser$Period

84
helen-invoice/src/main/java/com/devsoap/parsers/helen/Parser.java
View File

@ -6,17 +6,15 @@ import com.itextpdf.kernel.pdf.canvas.parser.PdfTextExtractor;
import java.io.FileNotFoundException;
import java.io.IOException;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.time.LocalDate;
import java.time.format.DateTimeFormatter;
import java.time.format.FormatStyle;
import java.time.format.TextStyle;
import java.util.HashMap;
import java.util.Locale;
import java.util.Objects;
import java.util.Scanner;
import java.util.function.Function;
import java.util.*;
import java.util.regex.Pattern;
import java.util.stream.Collectors;
public class Parser {
@ -31,46 +29,44 @@ public class Parser {
.ofLocalizedDate(FormatStyle.SHORT)
.withLocale(FI_LOCALE);
static class Period {
double basicPay = 0.0;
int dayEnergy = 0;
int nightEnergy = 0;
double dayEnergyEur = 0.0;
double nightEnergyEur = 0.0;
@Override
public String toString() {
return "Period{" +
"basicPay=" + basicPay +
", dayEnergy=" + dayEnergy +
", nightEnergy=" + nightEnergy +
", dayEnergyEur=" + dayEnergyEur +
", nightEnergyEur=" + nightEnergyEur +
'}';
}
public static class Period {
public double basicPay = 0.0;
public int dayEnergy = 0;
public int nightEnergy = 0;
public double dayEnergyEur = 0.0;
public double nightEnergyEur = 0.0;
}
public static void main(String[] args) {
var filename = args[0];
var daySiirtoKwh = 937;
var nightSiirtoKwh = 920;
var filename = Path.of(args[0]);
var daySiirtoKwh = Arrays
.stream(args[1].split(","))
.map(period -> period.split(":"))
.collect(Collectors.toMap(values -> values[0], values -> Integer.parseInt(values[1])));
var nightSiirtoKwh = Arrays
.stream(args[2].split(","))
.map(period -> period.split(":"))
.collect(Collectors.toMap(values -> values[0], values -> Integer.parseInt(values[1])));
var file = Paths.get(filename);
System.out.println("Kuukausi,Perusmaksu (energia),Perusmaksu (siirto),Päiväenergia (kWh),Päiväenergia " +
"(EUR),Yöenergia (kWh),Yöenergia (EUR),Päiväsiirto (kWh),Päiväsiirto (EUR),Yösiirto (kWh)" +
",Yösiirto (EUR),Vero");
parse(filename,daySiirtoKwh, nightSiirtoKwh).forEach((month,period ) -> {
var csv = String.format("%s,%.02f,,%d,%.02f,%d,%.02f,,,,", month,
period.dayEnergy, period.dayEnergyEur,period.nightEnergy, period.nightEnergyEur);
System.out.println(csv);
});
}
public static Map<String, Period> parse(Path file, Map<String, Integer> daySiirtoKwh,
Map<String, Integer> nightSiirtoKwh) {
try(var reader = new PdfReader(file.toFile())) {
var document = new PdfDocument(reader);
var page2 = document.getPage(2);
var text = PdfTextExtractor.getTextFromPage(page2);
var scanner = new Scanner(text);
System.out.println("Kuukausi,Perusmaksu (energia),Perusmaksu (siirto),Päiväenergia (kWh),Päiväenergia " +
"(EUR),Yöenergia (kWh),Yöenergia (EUR),Päiväsiirto (kWh),Päiväsiirto (EUR),Yösiirto (kWh)" +
",Yösiirto (EUR),Vero");
var periods = new HashMap<String, Period>();
while(scanner.hasNextLine()) {
var line = scanner.nextLine();
if(PERUSMAKSU_PATTERN.asPredicate().test(line)) {
@ -99,28 +95,22 @@ public class Parser {
var eurPerKwh = Double.parseDouble(matcher.group(4)
.replace(",", ".")) / 100.0;
var nightSiirto = nightSiirtoKwh.getOrDefault(month,0);
var daySriirto = daySiirtoKwh.getOrDefault(month,0);
periods.computeIfAbsent(month, s -> new Period());
periods.computeIfPresent(month, (s,p) -> {
p.dayEnergy = totalEnergy - nightSiirtoKwh;
p.dayEnergy = totalEnergy - nightSiirto;
p.dayEnergyEur = p.dayEnergy * eurPerKwh;
p.nightEnergy = totalEnergy - daySiirtoKwh;
p.nightEnergy = totalEnergy - daySriirto;
p.nightEnergyEur = p.nightEnergy * eurPerKwh;
return p;
});
}
}
}
periods.forEach((month,period ) -> {
var csv = String.format("%s,%.02f,,%d,%.02f,%d,%.02f,,,,", month, period.basicPay,
period.dayEnergy, period.dayEnergyEur,period.nightEnergy, period.nightEnergyEur);
System.out.println(csv);
});
} catch (FileNotFoundException e) {
e.printStackTrace();
return periods;
} catch (IOException e) {
e.printStackTrace();
throw new RuntimeException(e);
}
}
}

1
settings.gradle
View File

@ -1,4 +1,5 @@
rootProject.name = 'pdf-parsers'
include 'caruna-invoice'
include 'helen-invoice'
include 'composite-parsers'