kermitt2 / grobid / 385

Build # 385

Build Type

push

circleci

Committed by

Commit Message

review incremental training

Run Details

3 of 3 new or added lines in 2 files covered. (100.0%)

14846 of 37503 relevant lines covered (39.59%)

0.4 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

58.06

/grobid-core/src/main/java/org/grobid/core/jni/WapitiModel.java

package org.grobid.core.jni;

import fr.limsi.wapiti.SWIGTYPE_p_mdl_t;
import fr.limsi.wapiti.Wapiti;
import org.grobid.core.GrobidModel;
import org.grobid.core.GrobidModels;
import org.grobid.core.exceptions.GrobidException;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import java.io.File;

public class WapitiModel {
    public static final Logger LOGGER = LoggerFactory.getLogger(WapitiModel.class);

    private SWIGTYPE_p_mdl_t model;
    private File modelFile;

    public WapitiModel(File modelFile) {
        this.modelFile = modelFile;
        init();
    }

    public WapitiModel(GrobidModel grobidModel) {
        modelFile = new File(grobidModel.getModelPath());
        init();
    }

    private synchronized void init() {
        if (model != null) {
            return;
        }
        if (!modelFile.exists() || modelFile.isDirectory()) {
            throw new GrobidException("Model file does not exists or is a directory: " + modelFile.getAbsolutePath());
        }
        LOGGER.info("Loading model: " + modelFile + " (size: " + modelFile.length() + ")");
        model = WapitiWrapper.getModel(modelFile);
    }

    public String label(String data) {
        if (model == null) {
            LOGGER.warn("Model has been already closed, reopening: " + modelFile.getAbsolutePath());
            init();
        }
        String label = WapitiWrapper.label(model, data).trim();
        //TODO: VZ: Grobid currently expects tabs as separators whereas wapiti uses spaces for separating features.
        // for now it is safer to replace, although it does not look nice
        label = label.replaceAll(" ", "\t");
        return label;
    }

    public synchronized void close() {
        if (model != null) {
            Wapiti.freeModel(model);
            model = null;
        }
    }

    public static void train(File template, File trainingData, File outputModel) {
        train(template, trainingData, outputModel, "");
    }

    public static void train(File template, File trainingData, File outputModel, String params) {
                String args = String.format("train " + params + " -p %s %s %s", template.getAbsolutePath(), trainingData.getAbsolutePath(), outputModel.getAbsolutePath());
                //System.out.println("Training with equivalent command line: \n" + "wapiti " + args);
                Wapiti.runWapiti(args);
    }



}

1	package org.grobid.core.jni;
2
3	import fr.limsi.wapiti.SWIGTYPE_p_mdl_t;
4	import fr.limsi.wapiti.Wapiti;
5	import org.grobid.core.GrobidModel;
6	import org.grobid.core.GrobidModels;
7	import org.grobid.core.exceptions.GrobidException;
8	import org.slf4j.Logger;
9	import org.slf4j.LoggerFactory;
10
11	import java.io.File;
12
13	public class WapitiModel {
14	public static final Logger LOGGER = LoggerFactory.getLogger(WapitiModel.class);	1✔
15
16	private SWIGTYPE_p_mdl_t model;
17	private File modelFile;
18
19	public WapitiModel(File modelFile) {	×
20	this.modelFile = modelFile;	×
21	init();	×
22	}	×
23
24	public WapitiModel(GrobidModel grobidModel) {	1✔
25	modelFile = new File(grobidModel.getModelPath());	1✔
26	init();	1✔
27	}	1✔
28
29	private synchronized void init() {
30	if (model != null) {	1✔
31	return;	×
32	}
33	if (!modelFile.exists() \|\| modelFile.isDirectory()) {	1✔
34	throw new GrobidException("Model file does not exists or is a directory: " + modelFile.getAbsolutePath());	×
35	}
36	LOGGER.info("Loading model: " + modelFile + " (size: " + modelFile.length() + ")");	1✔
37	model = WapitiWrapper.getModel(modelFile);	1✔
38	}	1✔
39
40	public String label(String data) {
41	if (model == null) {	1✔
42	LOGGER.warn("Model has been already closed, reopening: " + modelFile.getAbsolutePath());	×
43	init();	×
44	}
45	String label = WapitiWrapper.label(model, data).trim();	1✔
46	//TODO: VZ: Grobid currently expects tabs as separators whereas wapiti uses spaces for separating features.
47	// for now it is safer to replace, although it does not look nice
48	label = label.replaceAll(" ", "\t");	1✔
49	return label;	1✔
50	}
51
52	public synchronized void close() {
53	if (model != null) {	1✔
54	Wapiti.freeModel(model);	1✔
55	model = null;	1✔
56	}
57	}	1✔
58
59	public static void train(File template, File trainingData, File outputModel) {
60	train(template, trainingData, outputModel, "");	×
61	}	×
62
63	public static void train(File template, File trainingData, File outputModel, String params) {
64	String args = String.format("train " + params + " -p %s %s %s", template.getAbsolutePath(), trainingData.getAbsolutePath(), outputModel.getAbsolutePath());	×
65	//System.out.println("Training with equivalent command line: \n" + "wapiti " + args);
66	Wapiti.runWapiti(args);	×
67	}	×
68
69
70
71	}

kermitt2 / grobid / 385

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous