2 * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3 * Copyright (C) $$Year-Rel$$ The Jalview Authors
5 * This file is part of Jalview.
7 * Jalview is free software: you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation, either version 3
10 * of the License, or (at your option) any later version.
12 * Jalview is distributed in the hope that it will be useful, but
13 * WITHOUT ANY WARRANTY; without even the implied warranty
14 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR
15 * PURPOSE. See the GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with Jalview. If not, see <http://www.gnu.org/licenses/>.
19 * The Jalview Authors are detailed in the 'AUTHORS' file.
24 * See https://issues.jalview.org/browse/JAL-4036
25 * The new Uniprot API is not dissimilar to the old one, but has some important changes.
26 * Some group names have changed slightly, some old groups have gone and there are quite a few new groups.
28 * Most changes are mappings of old column ids to new field ids. There are a handful of old
29 * columns not mapped to new fields, and new fields without an old column.
30 * [aside: not all possible columns were listed in the resources/fts/uniprot_data_columns.txt file.
31 * These were presumably additions after the file was created]
32 * For existing/mapped fields, the same preferences found in the resource file have been migrated to
33 * the new file with the new field name, id and group.
35 * The new mapped groups and files are stored and read from resources/fts/uniprot_data_columns-2022.txt.
37 * There is now no "sort" query string parameter.
39 * See https://www.uniprot.org/help/api_queries
41 * SIGNIFICANT CHANGE: Pagination is no longer performed using a record offset, but with a "cursor"
42 * query string parameter that is not really a cursor. The value is an opaque string that is passed (or
43 * rather a whole URL is passed) in the "Link" header of the HTTP response of the previous page.
44 * Where such a link is passed it is put into the cursors ArrayList.
45 * There are @Overridden methods in UniprotFTSPanel.
48 package jalview.fts.service.uniprot;
50 import java.lang.invoke.MethodHandles;
51 import java.net.MalformedURLException;
53 import java.util.ArrayList;
54 import java.util.Collection;
55 import java.util.List;
56 import java.util.Objects;
58 import javax.ws.rs.core.MediaType;
60 import com.sun.jersey.api.client.Client;
61 import com.sun.jersey.api.client.ClientResponse;
62 import com.sun.jersey.api.client.WebResource;
63 import com.sun.jersey.api.client.config.DefaultClientConfig;
65 import jalview.bin.Cache;
66 import jalview.bin.Console;
67 import jalview.fts.api.FTSData;
68 import jalview.fts.api.FTSDataColumnI;
69 import jalview.fts.core.FTSRestClient;
70 import jalview.fts.core.FTSRestRequest;
71 import jalview.fts.core.FTSRestResponse;
72 import jalview.util.ChannelProperties;
73 import jalview.util.MessageManager;
74 import jalview.util.Platform;
76 public class UniProtFTSRestClient extends FTSRestClient
78 private static final String DEFAULT_UNIPROT_DOMAIN = "https://rest.uniprot.org";
80 private static final String USER_AGENT = ChannelProperties
81 .getProperty("app_name", "Jalview") + " "
82 + Cache.getDefault("VERSION", "Unknown") + " "
83 + MethodHandles.lookup().lookupClass() + " help@jalview.org";
87 Platform.addJ2SDirectDatabaseCall(DEFAULT_UNIPROT_DOMAIN);
90 private static UniProtFTSRestClient instance = null;
92 public final String uniprotSearchEndpoint;
94 public UniProtFTSRestClient()
98 uniprotSearchEndpoint = Cache.getDefault("UNIPROT_DOMAIN",
99 DEFAULT_UNIPROT_DOMAIN) + "/uniprotkb/search";
102 @SuppressWarnings("unchecked")
104 public FTSRestResponse executeRequest(FTSRestRequest uniprotRestRequest)
107 return executeRequest(uniprotRestRequest, null);
110 public FTSRestResponse executeRequest(FTSRestRequest uniprotRestRequest,
111 String cursor) throws Exception
115 String wantedFields = getDataColumnsFieldsAsCommaDelimitedString(
116 uniprotRestRequest.getWantedFields());
117 int responseSize = (uniprotRestRequest.getResponseSize() == 0)
118 ? getDefaultResponsePageSize()
119 : uniprotRestRequest.getResponseSize();
121 int offSet = uniprotRestRequest.getOffSet();
123 if (isAdvancedQuery(uniprotRestRequest.getSearchTerm()))
125 query = uniprotRestRequest.getSearchTerm();
129 query = uniprotRestRequest.getFieldToSearchBy().equalsIgnoreCase(
130 "Search All") ? uniprotRestRequest.getSearchTerm()
132 // + uniprotRestRequest.getSearchTerm()
133 : uniprotRestRequest.getFieldToSearchBy() + ":"
134 + uniprotRestRequest.getSearchTerm();
137 // BH 2018 the trick here is to coerce the classes in Javascript to be
138 // different from the ones in Java yet still allow this to be correct for
141 Class<ClientResponse> clientResponseClass;
144 // JavaScript only -- coerce types to Java types for Java
145 client = (Client) (Object) new jalview.javascript.web.Client();
146 clientResponseClass = (Class<ClientResponse>) (Object) jalview.javascript.web.ClientResponse.class;
156 client = Client.create(new DefaultClientConfig());
157 clientResponseClass = ClientResponse.class;
160 WebResource webResource = null;
161 webResource = client.resource(uniprotSearchEndpoint)
162 .queryParam("format", "tsv")
163 .queryParam("fields", wantedFields)
164 .queryParam("size", String.valueOf(responseSize))
165 /* 2022 new api has no "sort"
166 * .queryParam("sort", "score")
168 .queryParam("query", query);
169 if (offSet != 0 && cursor != null && cursor.length() > 0)
170 // 2022 new api does not do pagination with an offset, it requires a
171 // "cursor" parameter with a key (given for the next page).
172 // (see https://www.uniprot.org/help/pagination)
174 webResource = webResource.queryParam("cursor", cursor);
177 "Uniprot FTS Request: " + webResource.getURI().toString());
178 // Execute the REST request
179 WebResource.Builder wrBuilder = webResource
180 .accept(MediaType.TEXT_PLAIN);
181 if (!Platform.isJS())
188 wrBuilder.header("User-Agent", USER_AGENT);
190 ClientResponse clientResponse = wrBuilder.get(clientResponseClass);
192 if (!Platform.isJS())
199 if (clientResponse.getHeaders().containsKey("Link"))
201 // extract the URL from the 'Link: <URL>; ref="stuff"' header
202 String linkHeader = clientResponse.getHeaders().get("Link")
204 if (linkHeader.indexOf("<") > -1)
206 String temp = linkHeader.substring(linkHeader.indexOf("<") + 1);
207 if (temp.indexOf(">") > -1)
209 String nextUrl = temp.substring(0, temp.indexOf(">"));
210 // then get the cursor value from the query string parameters
211 String nextCursor = getQueryParam("cursor", nextUrl);
212 setCursor(cursorPage + 1, nextCursor);
218 String uniProtTabDelimittedResponseString = clientResponse
219 .getEntity(String.class);
220 // Make redundant objects eligible for garbage collection to conserve
222 // System.out.println(">>>>> response : "
223 // + uniProtTabDelimittedResponseString);
224 if (clientResponse.getStatus() != 200)
226 String errorMessage = getMessageByHTTPStatusCode(
227 clientResponse.getStatus(), "Uniprot");
228 throw new Exception(errorMessage);
231 // new Uniprot API is not including a "X-Total-Results" header when there
233 List<String> resultsHeaders = clientResponse.getHeaders()
234 .get("X-Total-Results");
235 int xTotalResults = 0;
240 else if (resultsHeaders != null && resultsHeaders.size() >= 1)
242 xTotalResults = Integer.valueOf(resultsHeaders.get(0));
244 clientResponse = null;
246 return parseUniprotResponse(uniProtTabDelimittedResponseString,
247 uniprotRestRequest, xTotalResults);
248 } catch (Exception e)
250 Console.debug("Exception caught from response", e);
251 String exceptionMsg = e.getMessage();
252 if (exceptionMsg.contains("SocketException"))
254 // No internet connection
255 throw new Exception(MessageManager.getString(
256 "exception.unable_to_detect_internet_connection"));
258 else if (exceptionMsg.contains("UnknownHostException"))
260 // The server 'http://www.uniprot.org' is unreachable
261 throw new Exception(MessageManager.formatMessage(
262 "exception.fts_server_unreachable", "Uniprot"));
271 public boolean isAdvancedQuery(String query)
273 if (query.contains(" AND ") || query.contains(" OR ")
274 || query.contains(" NOT ") || query.contains(" ! ")
275 || query.contains(" || ") || query.contains(" && ")
276 || query.contains(":") || query.contains("-"))
283 public FTSRestResponse parseUniprotResponse(
284 String uniProtTabDelimittedResponseString,
285 FTSRestRequest uniprotRestRequest, int xTotalResults)
287 FTSRestResponse searchResult = new FTSRestResponse();
288 List<FTSData> result = null;
289 if (uniProtTabDelimittedResponseString == null
290 || uniProtTabDelimittedResponseString.trim().isEmpty())
292 searchResult.setNumberOfItemsFound(0);
295 String[] foundDataRow = uniProtTabDelimittedResponseString.split("\n");
296 if (foundDataRow != null && foundDataRow.length > 0)
298 result = new ArrayList<>();
299 boolean firstRow = true;
300 for (String dataRow : foundDataRow)
302 // The first data row is usually the header data. This should be
303 // filtered out from the rest of the data See: JAL-2485
309 // System.out.println(dataRow);
310 result.add(getFTSData(dataRow, uniprotRestRequest));
312 searchResult.setNumberOfItemsFound(xTotalResults);
313 searchResult.setSearchSummary(result);
319 // * Takes a collection of FTSDataColumnI and converts its 'code' values into
321 // * tab delimited string.
323 // * @param dataColumnFields
324 // * the collection of FTSDataColumnI to process
325 // * @return the generated comma delimited string from the supplied
326 // * FTSDataColumnI collection
328 // private String getDataColumnsFieldsAsTabDelimitedString(
329 // Collection<FTSDataColumnI> dataColumnFields)
331 // String result = "";
332 // if (dataColumnFields != null && !dataColumnFields.isEmpty())
334 // StringBuilder returnedFields = new StringBuilder();
335 // for (FTSDataColumnI field : dataColumnFields)
337 // if (field.getName().equalsIgnoreCase("Uniprot Id"))
339 // returnedFields.append("\t").append("Entry");
343 // returnedFields.append("\t").append(field.getName());
346 // returnedFields.deleteCharAt(0);
347 // result = returnedFields.toString();
352 public static FTSData getFTSData(String tabDelimittedDataStr,
353 FTSRestRequest request)
355 String primaryKey = null;
357 Object[] summaryRowData;
359 Collection<FTSDataColumnI> diplayFields = request.getWantedFields();
361 summaryRowData = new Object[diplayFields.size()];
362 String[] columns = tabDelimittedDataStr.split("\t");
363 for (FTSDataColumnI field : diplayFields)
367 String fieldData = columns[colCounter];
368 if (field.isPrimaryKeyColumn())
370 primaryKey = fieldData;
371 summaryRowData[colCounter++] = primaryKey;
373 else if (fieldData == null || fieldData.isEmpty())
375 summaryRowData[colCounter++] = null;
381 summaryRowData[colCounter++] = (field.getDataType()
382 .getDataTypeClass() == Integer.class)
383 ? Integer.valueOf(fieldData.replace(",", ""))
384 : (field.getDataType()
385 .getDataTypeClass() == Double.class)
386 ? Double.valueOf(fieldData)
388 } catch (Exception e)
391 System.out.println("offending value:" + fieldData);
394 } catch (Exception e)
396 // e.printStackTrace();
400 final String primaryKey1 = primaryKey;
402 final Object[] summaryRowData1 = summaryRowData;
406 public Object[] getSummaryData()
408 return summaryRowData1;
412 public Object getPrimaryKey()
418 * Returns a string representation of this object;
421 public String toString()
423 StringBuilder summaryFieldValues = new StringBuilder();
424 for (Object summaryField : summaryRowData1)
426 summaryFieldValues.append(
427 summaryField == null ? " " : summaryField.toString())
430 return summaryFieldValues.toString();
434 * Returns hash code value for this object
437 public int hashCode()
439 return Objects.hash(primaryKey1, this.toString());
443 public boolean equals(Object that)
445 return this.toString().equals(that.toString());
450 public static UniProtFTSRestClient getInstance()
452 if (instance == null)
454 instance = new UniProtFTSRestClient();
460 public String getColumnDataConfigFileName()
462 return "/fts/uniprot_data_columns-2022.txt";
465 /* 2022-07-20 bsoares
466 * used for the new API "cursor" pagination. See https://www.uniprot.org/help/pagination
468 private ArrayList<String> cursors;
470 private int cursorPage = 0;
472 protected int getCursorPage()
477 protected void setCursorPage(int i)
482 protected void setPrevCursorPage()
488 protected void setNextCursorPage()
493 protected void clearCursors()
495 cursors = new ArrayList(10);
498 protected String getCursor(int i)
500 return cursors.get(i);
503 protected String getNextCursor()
505 if (cursors.size() < cursorPage + 2)
507 return cursors.get(cursorPage + 1);
510 protected String getPrevCursor()
514 return cursors.get(cursorPage - 1);
517 protected void setCursor(int i, String c)
519 cursors.ensureCapacity(i + 1);
520 while (cursors.size() <= i)
526 "Set UniprotFRSRestClient cursors[" + i + "] to '" + c + "'");
530 public static String getQueryParam(String param, String u)
532 if (param == null || u == null)
536 URL url = new URL(u);
537 String[] kevs = url.getQuery().split("&");
538 for (int j = 0; j < kevs.length; j++)
540 String[] kev = kevs[j].split("=", 2);
541 if (param.equals(kev[0]))
546 } catch (MalformedURLException e)
548 // TODO Auto-generated catch block