EVA-3501 - adding new endpoint for parsing fasta file data (#80)

* adding new endpoint for parsing fasta file data

src/main/java/uk/ac/ebi/eva/evaseqcol/controller/admin/AdminController.java

@@ -10,6 +10,7 @@
 import org.springframework.http.ResponseEntity;
 import org.springframework.web.bind.annotation.PathVariable;
 import org.springframework.web.bind.annotation.PutMapping;
+import org.springframework.web.bind.annotation.RequestBody;
 import org.springframework.web.bind.annotation.RequestMapping;
 import org.springframework.web.bind.annotation.RestController;
 
@@ -76,4 +77,32 @@ public ResponseEntity<?> fetchAndInsertSeqColByAssemblyAccession(
             return new ResponseEntity<>(e.getMessage(), HttpStatus.CONFLICT);
         }
     }
+
+    @Operation(summary = "Add new sequence collection objects",
+            description = "Given FASTA file content, this endpoint will parse the content and use it to construct " +
+                    "seqCol objects with naming convention TEST and eventually save these seqCol objects into the database. " +
+                    "This is an authenticated endpoint, so it requires admin privileges to run it.")
+    @ApiResponses(value = {
+            @ApiResponse(responseCode = "201", description = "seqCol object(s) successfully inserted"),
+            @ApiResponse(responseCode = "409", description = "seqCol object(s) already exist(s)"),
+            @ApiResponse(responseCode = "404", description = "Assembly not found"),
+            @ApiResponse(responseCode = "400", description = "Bad request. (It can be a bad accession value)"),
+            @ApiResponse(responseCode = "500", description = "Server Error")
+    })
+    @PutMapping(value = "/seqcols/fasta/{accession}")
+    public ResponseEntity<?> fetchAndInsertSeqColByParsingFastaFile(@PathVariable(value = "accession") String accession, @RequestBody String fastaFileContent) {
+        try {
+            IngestionResultEntity ingestionResult = seqColService.fetchAndInsertAllSeqColInFastaFile(accession, fastaFileContent);
+            return new ResponseEntity<>(ingestionResult, HttpStatus.CREATED);
+        }  catch (IOException e) {
+            e.printStackTrace();
+            return new ResponseEntity<>(e.getMessage(), HttpStatus.INTERNAL_SERVER_ERROR);
+        } catch (DuplicateSeqColException e) {
+            return new ResponseEntity<>(e.getMessage(), HttpStatus.CONFLICT);
+        } catch (AssemblyNotFoundException e) {
+            return new ResponseEntity<>(e.getMessage(), HttpStatus.NOT_FOUND);
+        } catch (AssemblyAlreadyIngestedException e) {
+            return new ResponseEntity<>(e.getMessage(), HttpStatus.CONFLICT);
+        }
+    }
 }

src/main/java/uk/ac/ebi/eva/evaseqcol/datasource/NCBIAssemblySequenceDataSource.java

@@ -16,6 +16,7 @@
 import uk.ac.ebi.eva.evaseqcol.entities.AssemblySequenceEntity;
 import uk.ac.ebi.eva.evaseqcol.utils.GzipCompress;
 
+import java.io.ByteArrayInputStream;
 import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
@@ -43,6 +44,17 @@ public NCBIAssemblySequenceDataSource(NCBIBrowserFactory factory,
         this.readerFactory = readerFactory;
     }
 
+    public Optional<AssemblySequenceEntity> getAssemblySequencesByAccession(String accession, String fastaFileContent) throws IOException {
+        AssemblySequenceEntity assemblySequenceEntity;
+        try (InputStream stream = new ByteArrayInputStream(fastaFileContent.getBytes())) {
+            NCBIAssemblySequenceReader reader = readerFactory.build(stream, accession);
+            assemblySequenceEntity = reader.getAssemblySequencesEntity();
+            logger.info("FASTA file content with accession " + accession + " has been parsed successfully");
+        }
+
+        return Optional.of(assemblySequenceEntity);
+    }
+
     @Override
     public Optional<AssemblySequenceEntity> getAssemblySequencesByAccession(String accession) throws IOException, IllegalArgumentException {
         NCBIBrowser ncbiBrowser = factory.build();

src/main/java/uk/ac/ebi/eva/evaseqcol/datasource/NCBISeqColDataSource.java

@@ -16,6 +16,7 @@
 import uk.ac.ebi.eva.evaseqcol.utils.JSONLevelOne;
 
 import java.io.IOException;
+import java.util.Collections;
 import java.util.HashMap;
 import java.util.List;
 import java.util.Map;
@@ -86,4 +87,27 @@ public Optional<Map<String, Object>> getAllPossibleSeqColExtendedData(String acc
 
         return Optional.of(seqColResultData);
     }
+
+    public Optional<Map<String, Object>> getAllPossibleSeqColExtendedData(String accession, String fastaFileContent) throws IOException {
+        Map<String, Object> seqColResultData = new HashMap<>();
+
+        // Fetching Sequence Entity (FASTA File)
+        Optional<AssemblySequenceEntity> sequenceEntity = assemblySequenceDataSource.getAssemblySequencesByAccession(accession, fastaFileContent);
+        if (!sequenceEntity.isPresent()) {
+            logger.error("Could not parse FASTA file content: ");
+            return Optional.empty();
+        }
+        logger.info("FASTA file have been parsed successfully");
+
+        Map<String, Object> sameValueAttributesMap = new HashMap<>();
+        sameValueAttributesMap.put("extendedLengths", SeqColExtendedDataEntity.constructSeqColLengthsObject(sequenceEntity.get()));
+        sameValueAttributesMap.put("extendedSequences", SeqColExtendedDataEntity.constructSeqColSequencesObject(sequenceEntity.get()));
+        sameValueAttributesMap.put("extendedMd5Sequences", SeqColExtendedDataEntity.constructSeqColSequencesMd5Object(sequenceEntity.get()));
+
+        // Seqcol Result Data Map
+        seqColResultData.put("sameValueAttributes", sameValueAttributesMap);
+        seqColResultData.put("namesAttributes", Collections.singletonList(SeqColExtendedDataEntity
+                .constructSeqColNamesObjectWithRefSeqAndTESTNamingConvention(sequenceEntity.get())));
+        return Optional.of(seqColResultData);
+    }
 }

src/main/java/uk/ac/ebi/eva/evaseqcol/dus/NCBIAssemblySequenceReader.java

@@ -32,7 +32,7 @@ protected void parseFile() throws IOException, NullPointerException {
         while (line != null){
             if (line.startsWith(">")){
                 SeqColSequenceEntity sequence = new SeqColSequenceEntity();
-                String refSeq = line.substring(1, line.indexOf(' '));
+                String refSeq = line.substring(1).split(" ")[0];
                 sequence.setRefseq(refSeq);
                 line = reader.readLine();
                 StringBuilder sequenceValue = new StringBuilder();
@@ -45,6 +45,7 @@ protected void parseFile() throws IOException, NullPointerException {
                 String sha512Checksum = sha512ChecksumCalculator.calculateRefgetChecksum(sequenceValue.toString());
                 sequence.setSequenceMD5(md5checksum);
                 sequence.setSequence(sha512Checksum);
+                sequence.setLength(sequenceValue.length());
                 sequences.add(sequence);
             }
         }

src/main/java/uk/ac/ebi/eva/evaseqcol/entities/SeqColEntity.java

@@ -21,6 +21,6 @@ public abstract class SeqColEntity {
 
 
     public enum NamingConvention {
-        ENA, GENBANK, UCSC
+        ENA, GENBANK, UCSC, TEST
     }
 }

src/main/java/uk/ac/ebi/eva/evaseqcol/entities/SeqColExtendedDataEntity.java

@@ -25,6 +25,7 @@
 import java.util.Comparator;
 import java.util.LinkedList;
 import java.util.List;
+import java.util.stream.Collectors;
 
 @Entity
 @TypeDefs({
@@ -117,6 +118,20 @@ public static SeqColExtendedDataEntity<List<String>> constructSeqColNamesObjectB
         return seqColNamesObject;
     }
 
+    public static SeqColExtendedDataEntity<List<String>> constructSeqColNamesObjectWithRefSeqAndTESTNamingConvention(
+            AssemblySequenceEntity sequenceEntity) throws IOException {
+        SeqColExtendedDataEntity<List<String>> seqColNamesObject = new SeqColExtendedDataEntity<List<String>>().setAttributeType(
+                SeqColExtendedDataEntity.AttributeType.names);
+        seqColNamesObject.setNamingConvention(SeqColEntity.NamingConvention.TEST);
+        JSONExtData<List<String>> seqColNamesArray = new JSONStringListExtData();
+        List<String> namesList = sequenceEntity.getSequences().stream().map(s -> s.getRefseq()).collect(Collectors.toList());
+        DigestCalculator digestCalculator = new DigestCalculator();
+        seqColNamesArray.setObject(namesList);
+        seqColNamesObject.setExtendedSeqColData(seqColNamesArray);
+        seqColNamesObject.setDigest(digestCalculator.getSha512Digest(seqColNamesArray.toString()));
+        return seqColNamesObject;
+    }
+
     /**
      * Return the seqCol lengths array object*/
     public static SeqColExtendedDataEntity<List<Integer>> constructSeqColLengthsObject(AssemblyEntity assemblyEntity) throws IOException {
@@ -136,6 +151,21 @@ public static SeqColExtendedDataEntity<List<Integer>> constructSeqColLengthsObje
         return seqColLengthsObject;
     }
 
+
+    public static SeqColExtendedDataEntity<List<Integer>> constructSeqColLengthsObject(AssemblySequenceEntity sequenceEntity) throws IOException {
+        SeqColExtendedDataEntity<List<Integer>> seqColLengthsObject = new SeqColExtendedDataEntity<List<Integer>>().setAttributeType(
+                SeqColExtendedDataEntity.AttributeType.lengths);
+        JSONExtData<List<Integer>> seqColLengthsArray = new JSONIntegerListExtData();
+        List<Integer> lengthsList = sequenceEntity.getSequences().stream().map(s -> s.getLength()).collect(Collectors.toList());
+
+        DigestCalculator digestCalculator = new DigestCalculator();
+        seqColLengthsArray.setObject(lengthsList);
+        seqColLengthsObject.setExtendedSeqColData(seqColLengthsArray);
+        seqColLengthsObject.setDigest(digestCalculator.getSha512Digest(seqColLengthsArray.toString()));
+
+        return seqColLengthsObject;
+    }
+
     /**
      * Return the seqCol sequences array object*/
     public static SeqColExtendedDataEntity<List<String>> constructSeqColSequencesObject(

src/main/java/uk/ac/ebi/eva/evaseqcol/entities/SeqColSequenceEntity.java

@@ -11,6 +11,8 @@ public class SeqColSequenceEntity {
     private String sequenceMD5;
     @ApiModelProperty(value = "Sequence's defalut (ga4gh) checksum value")
     private String sequence;
+    @ApiModelProperty(value = "Sequence's length")
+    private Integer length;
 
     public SeqColSequenceEntity setRefseq(String refseq) {
         this.refseq = refseq;
@@ -26,4 +28,9 @@ public SeqColSequenceEntity setSequence(String sequence) {
         this.sequence = sequence;
         return this;
     }
+
+    public SeqColSequenceEntity setLength(Integer length) {
+        this.length = length;
+        return this;
+    }
 }

src/main/java/uk/ac/ebi/eva/evaseqcol/service/SeqColService.java

@@ -154,16 +154,24 @@ public void removeAllSeqCol() {
         extendedDataService.removeAllSeqColExtendedEntities();
     }
 
+    public IngestionResultEntity fetchAndInsertAllSeqColInFastaFile(String accession, String fastaFileContent) throws IOException {
+        Optional<Map<String, Object>> seqColDataMap = ncbiSeqColDataSource.getAllPossibleSeqColExtendedData(accession, fastaFileContent);
+        return createSeqColObjectsAndInsert(seqColDataMap, accession);
+    }
+
     /**
      * Fetch and insert all possible seqCol objects for the given assembly accession.
      * NOTE: All possible seqCol objects means with all possible/provided naming conventions that could be found in the
      * assembly report.
      * Return the list of level 0 digests of the inserted seqcol objects*/
-    public IngestionResultEntity fetchAndInsertAllSeqColByAssemblyAccession(
-            String assemblyAccession) throws IOException, DuplicateSeqColException, AssemblyNotFoundException,
-            AssemblyAlreadyIngestedException{
-        Optional<Map<String, Object>> seqColDataMap = ncbiSeqColDataSource
-                .getAllPossibleSeqColExtendedData(assemblyAccession);
+    public IngestionResultEntity fetchAndInsertAllSeqColByAssemblyAccession(String assemblyAccession) throws IOException {
+        Optional<Map<String, Object>> seqColDataMap = ncbiSeqColDataSource.getAllPossibleSeqColExtendedData(assemblyAccession);
+        return createSeqColObjectsAndInsert(seqColDataMap, assemblyAccession);
+    }
+
+
+    public IngestionResultEntity createSeqColObjectsAndInsert(Optional<Map<String, Object>> seqColDataMap,
+                                                              String assemblyAccession) throws IOException {
         if (!seqColDataMap.isPresent()) {
             logger.warn("No seqCol data corresponding to assemblyAccession " + assemblyAccession + " could be found on NCBI datasource");
             throw new AssemblyNotFoundException(assemblyAccession);