Revision 52931
Added by Claudio Atzori over 5 years ago
ScholexplorerMapper.java | ||
---|---|---|
1 | 1 |
package eu.dnetlib.data.mapreduce.hbase.dataimport; |
2 | 2 |
|
3 |
import java.io.IOException; |
|
4 |
import java.util.HashMap; |
|
5 |
import java.util.List; |
|
6 |
import java.util.Map; |
|
3 | 7 |
|
4 | 8 |
import com.google.gson.JsonObject; |
5 | 9 |
import com.google.gson.JsonParser; |
... | ... | |
7 | 11 |
import eu.dnetlib.actionmanager.actions.AtomicAction; |
8 | 12 |
import eu.dnetlib.actionmanager.common.Agent; |
9 | 13 |
import eu.dnetlib.miscutils.datetime.DateUtils; |
14 |
import org.apache.hadoop.io.LongWritable; |
|
10 | 15 |
import org.apache.hadoop.io.Text; |
11 | 16 |
import org.apache.hadoop.mapreduce.Mapper; |
12 | 17 |
|
13 |
import java.io.IOException; |
|
14 |
import java.util.HashMap; |
|
15 |
import java.util.List; |
|
16 |
import java.util.Map; |
|
18 |
public class ScholexplorerMapper extends Mapper<LongWritable, Text, Text, Text> { |
|
17 | 19 |
|
18 |
public class ScholexplorerMapper extends Mapper<Text, Text, Text, Text> { |
|
19 |
|
|
20 | 20 |
private ActionFactory factory; |
21 | 21 |
private JsonParser parser; |
22 | 22 |
private String setName; |
... | ... | |
27 | 27 |
private String dateOfCollection; |
28 | 28 |
private Text keyout; |
29 | 29 |
private Text valueOut; |
30 |
private Map<String, ScholExplorerConfiguration> configurationMap= new HashMap<>();
|
|
30 |
private Map<String, ScholExplorerConfiguration> conf = new HashMap<>();
|
|
31 | 31 |
|
32 | 32 |
@Override |
33 | 33 |
protected void setup(Context context) throws IOException, InterruptedException { |
34 | 34 |
factory = new ActionFactory(); |
35 | 35 |
parser = new JsonParser(); |
36 |
configurationMap.put("issn", new ScholExplorerConfiguration(null, false));
|
|
37 |
configurationMap.put("pmid", new ScholExplorerConfiguration("pmid", true,"https://www.ncbi.nlm.nih.gov/pubmed/%s"));
|
|
38 |
configurationMap.put("doi", new ScholExplorerConfiguration("doi", true,"http://dx.doi.org/%s"));
|
|
39 |
configurationMap.put("pbmid", new ScholExplorerConfiguration("pmid", true,"https://www.ncbi.nlm.nih.gov/pubmed/%s"));
|
|
40 |
configurationMap.put("openaire ", new ScholExplorerConfiguration(null, false));
|
|
41 |
configurationMap.put("pmcid", new ScholExplorerConfiguration("pmc", true,"https://europepmc.org/articles/%s"));
|
|
42 |
configurationMap.put("pubmedid", new ScholExplorerConfiguration("pmid", true,"https://www.ncbi.nlm.nih.gov/pubmed/%s"));
|
|
43 |
configurationMap.put("icpsr", new ScholExplorerConfiguration(null, false));
|
|
44 |
configurationMap.put("dnet", new ScholExplorerConfiguration(null, false));
|
|
45 |
configurationMap.put("url ", new ScholExplorerConfiguration(null, true,"%s"));
|
|
46 |
configurationMap.put("openaire", new ScholExplorerConfiguration(null, false));
|
|
36 |
conf.put("issn", new ScholExplorerConfiguration(null, false)); |
|
37 |
conf.put("pmid", new ScholExplorerConfiguration("pmid", true,"https://www.ncbi.nlm.nih.gov/pubmed/%s")); |
|
38 |
conf.put("doi", new ScholExplorerConfiguration("doi", true,"http://dx.doi.org/%s")); |
|
39 |
conf.put("pbmid", new ScholExplorerConfiguration("pmid", true,"https://www.ncbi.nlm.nih.gov/pubmed/%s")); |
|
40 |
conf.put("openaire ", new ScholExplorerConfiguration(null, false)); |
|
41 |
conf.put("pmcid", new ScholExplorerConfiguration("pmc", true,"https://europepmc.org/articles/%s")); |
|
42 |
conf.put("pubmedid", new ScholExplorerConfiguration("pmid", true,"https://www.ncbi.nlm.nih.gov/pubmed/%s")); |
|
43 |
conf.put("icpsr", new ScholExplorerConfiguration(null, false)); |
|
44 |
conf.put("dnet", new ScholExplorerConfiguration(null, false)); |
|
45 |
conf.put("url ", new ScholExplorerConfiguration(null, true,"%s")); |
|
46 |
conf.put("openaire", new ScholExplorerConfiguration(null, false)); |
|
47 | 47 |
setName = context.getConfiguration().get("setName"); |
48 | 48 |
agent= new Agent(context.getConfiguration().get("agentId"), context.getConfiguration().get("agentName"), Agent.AGENT_TYPE.service); |
49 | 49 |
nsPrefix = context.getConfiguration().get("ns_prefix"); |
50 | 50 |
dsName = context.getConfiguration().get("dsName"); |
51 | 51 |
dsId = context.getConfiguration().get("dsId"); |
52 | 52 |
dateOfCollection = context.getConfiguration().get("dateOfCollection", DateUtils.now_ISO8601()); |
53 |
|
|
53 | 54 |
keyout = new Text(""); |
54 | 55 |
valueOut = new Text(""); |
55 |
|
|
56 |
|
|
57 | 56 |
} |
58 | 57 |
|
59 | 58 |
@Override |
60 |
protected void map(Text key, Text value, Context context) throws IOException, InterruptedException {
|
|
59 |
protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
|
|
61 | 60 |
|
62 |
final String inputJson =value.toString(); |
|
61 |
final String inputJson = value.toString();
|
|
63 | 62 |
final JsonObject rootElement = parser.parse(inputJson).getAsJsonObject(); |
64 |
final List<AtomicAction> actions = ScholixToActions.generateActionsFromScholix(rootElement, configurationMap, setName,agent, factory, nsPrefix, dsName, dsId, dateOfCollection);
|
|
63 |
final List<AtomicAction> actions = ScholixToActions.generateActionsFromScholix(rootElement, conf, setName,agent, factory, nsPrefix, dsName, dsId, dateOfCollection); |
|
65 | 64 |
for(final AtomicAction action : actions) { |
66 | 65 |
keyout.set(action.getRowKey()); |
67 | 66 |
valueOut.set(action.toJSON()); |
... | ... | |
69 | 68 |
} |
70 | 69 |
} |
71 | 70 |
|
72 |
|
|
73 |
|
|
74 |
|
|
75 |
|
|
76 |
|
|
77 |
|
|
78 |
|
|
79 | 71 |
} |
Also available in: Unified diff
small fixes