001/**
002 * The contents of this file are subject to the Mozilla Public License Version 1.1
003 * (the "License"); you may not use this file except in compliance with the License.
004 * You may obtain a copy of the License at http://www.mozilla.org/MPL/
005 * Software distributed under the License is distributed on an "AS IS" basis,
006 * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License for the
007 * specific language governing rights and limitations under the License.
008 *
009 * The Original Code is "PipeParser.java".  Description:
010 * "An implementation of Parser that supports traditionally encoded (i.e"
011 *
012 * The Initial Developer of the Original Code is University Health Network. Copyright (C)
013 * 2001.  All Rights Reserved.
014 *
015 * Contributor(s): Kenneth Beaton.
016 *
017 * Alternatively, the contents of this file may be used under the terms of the
018 * GNU General Public License (the  "GPL"), in which case the provisions of the GPL are
019 * applicable instead of those above.  If you wish to allow use of your version of this
020 * file only under the terms of the GPL and not to allow others to use your version
021 * of this file under the MPL, indicate your decision by deleting  the provisions above
022 * and replace  them with the notice and other provisions required by the GPL License.
023 * If you do not delete the provisions above, a recipient may use your version of
024 * this file under either the MPL or the GPL.
025 *
026 */
027
028package ca.uhn.hl7v2.parser;
029
030import java.util.ArrayList;
031import java.util.Arrays;
032import java.util.HashMap;
033import java.util.List;
034import java.util.Map;
035import java.util.Set;
036import java.util.StringTokenizer;
037
038import ca.uhn.hl7v2.validation.ValidationContext;
039import org.slf4j.Logger;
040import org.slf4j.LoggerFactory;
041
042import ca.uhn.hl7v2.DefaultHapiContext;
043import ca.uhn.hl7v2.ErrorCode;
044import ca.uhn.hl7v2.HL7Exception;
045import ca.uhn.hl7v2.HapiContext;
046import ca.uhn.hl7v2.Version;
047import ca.uhn.hl7v2.model.AbstractSuperMessage;
048import ca.uhn.hl7v2.model.DoNotCacheStructure;
049import ca.uhn.hl7v2.model.Group;
050import ca.uhn.hl7v2.model.Message;
051import ca.uhn.hl7v2.model.Primitive;
052import ca.uhn.hl7v2.model.Segment;
053import ca.uhn.hl7v2.model.Structure;
054import ca.uhn.hl7v2.model.SuperStructure;
055import ca.uhn.hl7v2.model.Type;
056import ca.uhn.hl7v2.model.Varies;
057import ca.uhn.hl7v2.util.ReflectionUtil;
058import ca.uhn.hl7v2.util.Terser;
059import ca.uhn.hl7v2.validation.impl.NoValidation;
060import ca.uhn.hl7v2.validation.impl.ValidationContextFactory;
061
062/**
063 * An implementation of Parser that supports traditionally encoded (ie delimited
064 * with characters like |, ^, and ~) HL7 messages. Unexpected segments and
065 * fields are parsed into generic elements that are added to the message.
066 * 
067 * @see ParserConfiguration for configuration options which may affect parser encoding and decoding behaviour
068 * @author Bryan Tripp (bryan_tripp@sourceforge.net)
069 */
070public class PipeParser extends Parser {
071
072        private static final Logger log = LoggerFactory.getLogger(PipeParser.class);
073
074        /**
075         * The HL7 ER7 segment delimiter (see section 2.8 of spec)
076         */
077        final static String SEGMENT_DELIMITER = "\r";
078
079        private final HashMap<Class<? extends Message>, HashMap<String, StructureDefinition>> myStructureDefinitions = new HashMap<>();
080
081        /**
082         * System property key. If value is "true", legacy mode will default to true
083         * 
084         * @see #isLegacyMode()
085         * @deprecated This will be removed in HAPI 3.0
086         */
087        public static final String DEFAULT_LEGACY_MODE_PROPERTY = "ca.uhn.hl7v2.parser.PipeParser.default_legacy_mode";
088
089        private Boolean myLegacyMode = null;
090
091        public PipeParser() {
092                super();
093        }
094
095        /**
096         * @param context
097         *            the context containing all configuration items to be used
098         */
099        public PipeParser(HapiContext context) {
100                super(context);
101        }
102
103        /**
104         * Creates a new PipeParser
105         * 
106         * @param theFactory
107         *            custom factory to use for model class lookup
108         */
109        public PipeParser(ModelClassFactory theFactory) {
110                super(theFactory);
111        }
112
113    @Override
114    public void setValidationContext(ValidationContext context) {
115        super.setValidationContext(context);
116    }
117
118    /**
119         * Returns a String representing the encoding of the given message, if the
120         * encoding is recognized. For example if the given message appears to be
121         * encoded using HL7 2.x XML rules then "XML" would be returned. If the
122         * encoding is not recognized then null is returned. That this method
123         * returns a specific encoding does not guarantee that the message is
124         * correctly encoded (e.g. well formed XML) - just that it is not encoded
125         * using any other encoding than the one returned.
126         */
127        public String getEncoding(String message) {
128                return EncodingDetector.isEr7Encoded(message) ? getDefaultEncoding() : null;
129        }
130
131        /**
132         * @return the preferred encoding of this Parser
133         */
134        public String getDefaultEncoding() {
135                return "VB";
136        }
137
138        /**
139         * @deprecated this method should not be public
140         * @param message HL7 message
141         * @return message structure
142         * @throws HL7Exception
143         */
144        public String getMessageStructure(String message) throws HL7Exception {
145                return getStructure(message).messageStructure;
146        }
147
148        /**
149         * @return the message structure from MSH-9-3
150         */
151        private MessageStructure getStructure(String message) throws HL7Exception {
152                EncodingCharacters ec = getEncodingChars(message);
153                String messageStructure;
154                boolean explicityDefined = true;
155                String wholeFieldNine;
156                try {
157                        String[] fields = split(message.substring(0, Math.max(message.indexOf(SEGMENT_DELIMITER), message.length())), String.valueOf(ec.getFieldSeparator()));
158                        wholeFieldNine = fields[8];
159
160                        // message structure is component 3 but we'll accept a composite of
161                        // 1 and 2 if there is no component 3 ...
162                        // if component 1 is ACK, then the structure is ACK regardless of
163                        // component 2
164                        String[] comps = split(wholeFieldNine, String.valueOf(ec.getComponentSeparator()));
165                        if (comps.length >= 3) {
166                                messageStructure = comps[2];
167                        } else if (comps.length > 0 && comps[0] != null && comps[0].equals("ACK")) {
168                                messageStructure = "ACK";
169                        } else if (comps.length == 2) {
170                                explicityDefined = false;
171                                messageStructure = comps[0] + "_" + comps[1];
172                        }
173                        /*
174                         * else if (comps.length == 1 && comps[0] != null &&
175                         * comps[0].equals("ACK")) { messageStructure = "ACK"; //it's common
176                         * for people to only populate component 1 in an ACK msg }
177                         */
178                        else {
179                                String buf = "Can't determine message structure from MSH-9: " + wholeFieldNine +
180                                                " HINT: there are only " +
181                                                comps.length +
182                                                " of 3 components present";
183                                throw new HL7Exception(buf, ErrorCode.UNSUPPORTED_MESSAGE_TYPE);
184                        }
185                } catch (IndexOutOfBoundsException e) {
186                        throw new HL7Exception("Can't find message structure (MSH-9-3): " + e.getMessage(), ErrorCode.UNSUPPORTED_MESSAGE_TYPE);
187                }
188
189                return new MessageStructure(messageStructure, explicityDefined);
190        }
191
192        /**
193         * Returns object that contains the field separator and encoding characters
194         * for this message.
195         *
196         * There's an additional character starting with v2.7 (truncation), but we will
197         * accept it in messages of any version.
198         * 
199         * @throws HL7Exception
200         */
201        private static EncodingCharacters getEncodingChars(String message) throws HL7Exception {
202                if (message.length() < 9) {
203                        throw new HL7Exception("Invalid message content: \"" + message + "\"");
204                }
205                return new EncodingCharacters(message.charAt(3), message.substring(4, 9));
206        }
207
208        /**
209         * Parses a message string and returns the corresponding Message object.
210         * Unexpected segments added at the end of their group.
211         * 
212         * @throws HL7Exception
213         *             if the message is not correctly formatted.
214         * @throws EncodingNotSupportedException
215         *             if the message encoded is not supported by this parser.
216         */
217        protected Message doParse(String message, String version) throws HL7Exception {
218
219                // try to instantiate a message object of the right class
220                MessageStructure structure = getStructure(message);
221                Message m = instantiateMessage(structure.messageStructure, version, structure.explicitlyDefined);
222                m.setParser(this);
223                parse(m, message);
224                return m;
225        }
226
227        /**
228         * {@inheritDoc}
229         */
230        protected Message doParseForSpecificPackage(String message, String version, String packageName) throws HL7Exception {
231
232                // try to instantiate a message object of the right class
233                MessageStructure structure = getStructure(message);
234                Message m = instantiateMessageInASpecificPackage(structure.messageStructure, version, structure.explicitlyDefined, packageName);
235
236                parse(m, message);
237
238                return m;
239        }
240
241        /**
242         * Generates (or returns the cached value of) the message
243         */
244        private IStructureDefinition getStructureDefinition(Message theMessage) throws HL7Exception {
245
246                Class<? extends Message> clazz = theMessage.getClass();
247                HashMap<String, StructureDefinition> definitions = myStructureDefinitions.get(clazz);
248
249                StructureDefinition retVal;
250                if (definitions != null) {
251                        retVal = definitions.get(theMessage.getName());
252                        if (retVal != null) {
253                                return retVal;
254                        }
255                }
256
257                if (theMessage instanceof SuperStructure) {
258                        Set<String> appliesTo = ((SuperStructure) theMessage).getStructuresWhichChildAppliesTo("MSH");
259                        if (!appliesTo.contains(theMessage.getName())) {
260                                throw new HL7Exception("Superstructure " + theMessage.getClass().getSimpleName() + " does not apply to message " + theMessage.getName() + ", can not parse.");
261                        }
262                }
263                
264                if (clazz.isAnnotationPresent(DoNotCacheStructure.class)) {
265                        Holder<StructureDefinition> previousLeaf = new Holder<>();
266                        retVal = createStructureDefinition(theMessage, previousLeaf, theMessage.getName());
267                } else {
268                        Message message = ReflectionUtil.instantiateMessage(clazz, getFactory());
269                        Holder<StructureDefinition> previousLeaf = new Holder<>();
270                        retVal = createStructureDefinition(message, previousLeaf, theMessage.getName());
271
272                        if (!myStructureDefinitions.containsKey(clazz)) {
273                                myStructureDefinitions.put(clazz, new HashMap<>());
274                        }
275                        myStructureDefinitions.get(clazz).put(theMessage.getName(), retVal);
276                }
277
278                return retVal;
279        }
280
281        private StructureDefinition createStructureDefinition(Structure theStructure, Holder<StructureDefinition> thePreviousLeaf, String theStructureName) throws HL7Exception {
282
283                StructureDefinition retVal = new StructureDefinition();
284                retVal.setName(theStructure.getName());
285
286                if (theStructure instanceof Group) {
287                        retVal.setSegment(false);
288                        Group group = (Group) theStructure;
289                        int index = 0;
290                        List<String> childNames = Arrays.asList(group.getNames());
291                        
292                        /*
293                         * For SuperStructures, which can hold more than one type of structure,
294                         * we only actually bring in the child names that are actually a part
295                         * of the structure we are trying to parse
296                         */
297                        if (theStructure instanceof SuperStructure) {
298                                String struct = theStructureName;
299                                Map<String, String> evtMap = new DefaultModelClassFactory().getEventMapForVersion(Version.versionOf(theStructure.getMessage().getVersion()));
300                                if (evtMap.containsKey(struct)) {
301                                        struct = evtMap.get(struct);
302                                }
303                                childNames = ((SuperStructure) theStructure).getChildNamesForStructure(struct);
304                        }
305                        
306                        for (String nextName : childNames) {
307                                Structure nextChild = group.get(nextName);
308                                StructureDefinition structureDefinition = createStructureDefinition(nextChild, thePreviousLeaf, theStructureName);
309                                structureDefinition.setNameAsItAppearsInParent(nextName);
310                                structureDefinition.setRepeating(group.isRepeating(nextName));
311                                structureDefinition.setRequired(group.isRequired(nextName));
312                                structureDefinition.setChoiceElement(group.isChoiceElement(nextName));
313                                structureDefinition.setPosition(index++);
314                                structureDefinition.setParent(retVal);
315                                retVal.addChild(structureDefinition);
316                        }
317                } else {
318                        if (thePreviousLeaf.getObject() != null) {
319                                thePreviousLeaf.getObject().setNextLeaf(retVal);
320                        }
321                        thePreviousLeaf.setObject(retVal);
322                        retVal.setSegment(true);
323                }
324
325                return retVal;
326        }
327
328        /**
329         * Parses a segment string and populates the given Segment object.
330         * Unexpected fields are added as Varies' at the end of the segment.
331     *
332         * @param destination segment to parse the segment string into
333     * @param segment encoded segment
334     * @param encodingChars encoding characters to be used
335         * @throws HL7Exception
336         *             if the given string does not contain the given segment or if
337         *             the string is not encoded properly
338         */
339        public void parse(Segment destination, String segment, EncodingCharacters encodingChars) throws HL7Exception {
340                parse(destination, segment, encodingChars, 0);
341        }
342
343        /**
344         * Parses a segment string and populates the given Segment object.
345         * Unexpected fields are added as Varies' at the end of the segment.
346         *
347     * @param destination segment to parse the segment string into
348     * @param segment encoded segment
349     * @param encodingChars encoding characters to be used
350         * @param theRepetition the repetition number of this segment within its group
351         * @throws HL7Exception
352         *             if the given string does not contain the given segment or if
353         *             the string is not encoded properly
354         */
355        public void parse(Segment destination, String segment, EncodingCharacters encodingChars, int theRepetition) throws HL7Exception {
356                int fieldOffset = 0;
357                if (isDelimDefSegment(destination.getName())) {
358                        fieldOffset = 1;
359                        // set field 1 to fourth character of string
360                        Terser.set(destination, 1, 0, 1, 1, String.valueOf(encodingChars.getFieldSeparator()));
361                }
362
363                String[] fields = split(segment, String.valueOf(encodingChars.getFieldSeparator()));
364                // destination.setName(fields[0]);
365                for (int i = 1; i < fields.length; i++) {
366                        String[] reps = split(fields[i], String.valueOf(encodingChars.getRepetitionSeparator()));
367
368                        // MSH-2 will get split incorrectly so we have to fudge it ...
369                        boolean isMSH2 = isDelimDefSegment(destination.getName()) && i + fieldOffset == 2;
370                        if (isMSH2) {
371                                reps = new String[1];
372                                reps[0] = fields[i];
373                        }
374
375                        for (int j = 0; j < reps.length; j++) {
376                                try {
377                                        log.trace("Parsing field {} repetition {}", i + fieldOffset, j);
378                                        Type field = destination.getField(i + fieldOffset, j);
379                                        if (isMSH2) {
380                                                Terser.getPrimitive(field, 1, 1).setValue(reps[j]);
381                                        } else {
382                                                parse(field, reps[j], encodingChars);
383                                        }
384                                } catch (HL7Exception e) {
385                                        // set the field location and throw again ...
386                                        e.setFieldPosition(i);
387                                        if (theRepetition > 1) {
388                                                e.setSegmentRepetition(theRepetition);
389                                        }
390                                        e.setSegmentName(destination.getName());
391                                        throw e;
392                                }
393                        }
394                }
395
396                // set data type of OBX-5
397                if (destination.getClass().getName().contains("OBX")) {
398                        FixFieldDataType.fixOBX5(destination, getFactory(), getHapiContext().getParserConfiguration());
399                }
400        // set data type of MFE-4
401        if (destination.getClass().getName().contains("MFE") &&
402                Version.versionOf(destination.getMessage().getVersion()).isGreaterThan(Version.V23)) {
403            FixFieldDataType.fixMFE4(destination, getFactory(), getHapiContext().getParserConfiguration());
404        }
405
406        }
407
408        /**
409         * @return true if the segment is MSH, FHS, or BHS. These need special
410         *         treatment because they define delimiters.
411         * @param theSegmentName
412         *            segment name
413         */
414        private static boolean isDelimDefSegment(String theSegmentName) {
415                boolean is = false;
416                if (theSegmentName.equals("MSH") || theSegmentName.equals("FHS") || theSegmentName.equals("BHS")) {
417                        is = true;
418                }
419                return is;
420        }
421
422        /**
423         * Fills a field with values from an unparsed string representing the field.
424         * 
425         * @param destinationField
426         *            the field Type
427         * @param data
428         *            the field string (including all components and subcomponents;
429         *            not including field delimiters)
430         * @param encodingCharacters
431         *            the encoding characters used in the message
432         */
433        @Override
434        public void parse(Type destinationField, String data, EncodingCharacters encodingCharacters) throws HL7Exception {
435                String[] components = split(data, String.valueOf(encodingCharacters.getComponentSeparator()));
436                for (int i = 0; i < components.length; i++) {
437                        String[] subcomponents = split(components[i], String.valueOf(encodingCharacters.getSubcomponentSeparator()));
438                        for (int j = 0; j < subcomponents.length; j++) {
439                                String val = subcomponents[j];
440                                if (val != null) {
441                                        val = getParserConfiguration().getEscaping().unescape(val, encodingCharacters);
442                                }
443                                Terser.getPrimitive(destinationField, i + 1, j + 1).setValue(val);
444                        }
445                }
446        }
447
448        /**
449         * Splits the given composite string into an array of components using the
450         * given delimiter.
451     *
452     * @param composite encoded composite string
453     * @param delim delimiter to split upon
454     * @return split string
455         */
456        public static String[] split(String composite, String delim) {
457                ArrayList<String> components = new ArrayList<>();
458
459                // defend against evil nulls
460                if (composite == null)
461                        composite = "";
462                if (delim == null)
463                        delim = "";
464
465                StringTokenizer tok = new StringTokenizer(composite, delim, true);
466                boolean previousTokenWasDelim = true;
467                while (tok.hasMoreTokens()) {
468                        String thisTok = tok.nextToken();
469                        if (thisTok.equals(delim)) {
470                                if (previousTokenWasDelim)
471                                        components.add(null);
472                                previousTokenWasDelim = true;
473                        } else {
474                                components.add(thisTok);
475                                previousTokenWasDelim = false;
476                        }
477                }
478
479                String[] ret = new String[components.size()];
480                for (int i = 0; i < components.size(); i++) {
481                        ret[i] = components.get(i);
482                }
483
484                return ret;
485        }
486
487        /**
488         * {@inheritDoc }
489         */
490        @Override
491        public String doEncode(Segment structure, EncodingCharacters encodingCharacters) {
492                return encode(structure, encodingCharacters, getParserConfiguration(), null);
493        }
494
495        /**
496         * {@inheritDoc }
497         */
498        @Override
499        public String doEncode(Type type, EncodingCharacters encodingCharacters) {
500                return encode(type, encodingCharacters, getParserConfiguration(), null);
501        }
502
503        /**
504         * Encodes the given Type, using the given encoding characters. It is
505         * assumed that the Type represents a complete field rather than a
506         * component.
507     *
508     * @param source type to be encoded
509     * @param encodingChars encoding characters to be used
510     * @return encoded type
511         */
512        public static String encode(Type source, EncodingCharacters encodingChars) {
513                return encode(source, encodingChars, source.getMessage().getParser().getParserConfiguration(), null);
514        }
515
516        private static String encode(Type source, EncodingCharacters encodingChars, ParserConfiguration parserConfig, String currentTerserPath) {
517                if (source instanceof Varies) {
518                        Varies varies = (Varies) source;
519                        if (varies.getData() != null) {
520                                source = varies.getData();
521                        }
522                }
523
524                StringBuilder field = new StringBuilder();
525                for (int i = 1; i <= Terser.numComponents(source); i++) {
526                        StringBuilder comp = new StringBuilder();
527                        for (int j = 1; j <= Terser.numSubComponents(source, i); j++) {
528                                Primitive p = Terser.getPrimitive(source, i, j);
529                                comp.append(encodePrimitive(p, parserConfig.getEscaping(), encodingChars));
530                                comp.append(encodingChars.getSubcomponentSeparator());
531                        }
532                        field.append(stripExtraDelimiters(comp.toString(), encodingChars.getSubcomponentSeparator()));
533                        field.append(encodingChars.getComponentSeparator());
534                }
535
536                int forceUpToFieldNum = 0;
537                if (parserConfig != null && currentTerserPath != null) {
538                        for (String nextPath : parserConfig.getForcedEncode()) {
539                                if (nextPath.startsWith(currentTerserPath + "-") && nextPath.length() > currentTerserPath.length()) {
540                                        int endOfFieldDef = nextPath.indexOf('-', currentTerserPath.length());
541                                        if (endOfFieldDef == -1) {
542                                                forceUpToFieldNum = 0;
543                                                break;
544                                        }
545                                        String fieldNumString = nextPath.substring(endOfFieldDef + 1);
546                                        if (fieldNumString.length() > 0) {
547                                                forceUpToFieldNum = Math.max(forceUpToFieldNum, Integer.parseInt(fieldNumString));
548                                        }
549                                }
550                        }
551                }
552
553                char componentSeparator = encodingChars.getComponentSeparator();
554                String retVal = stripExtraDelimiters(field.toString(), componentSeparator);
555
556                while (forceUpToFieldNum > 0 && (countInstancesOf(retVal, componentSeparator) + 1) < forceUpToFieldNum) {
557                        retVal = retVal + componentSeparator;
558                }
559
560                return retVal;
561        }
562
563        private static String encodePrimitive(Primitive p, Escaping escaping, EncodingCharacters encodingChars) {
564                String val = (p).getValue();
565                if (val == null) {
566                        val = "";
567                } else {
568                        val = escaping.escape(val, encodingChars);
569                }
570                return val;
571        }
572
573        /**
574         * Removes unecessary delimiters from the end of a field or segment. This
575         * seems to be more convenient than checking to see if they are needed while
576         * we are building the encoded string.
577         */
578        private static String stripExtraDelimiters(String in, char delim) {
579                char[] chars = in.toCharArray();
580
581                // search from back end for first occurance of non-delimiter ...
582                int c = chars.length - 1;
583                boolean found = false;
584                while (c >= 0 && !found) {
585                        if (chars[c--] != delim)
586                                found = true;
587                }
588
589                String ret = "";
590                if (found)
591                        ret = String.valueOf(chars, 0, c + 2);
592                return ret;
593        }
594
595        /**
596         * Formats a Message object into an HL7 message string using the given
597         * encoding.
598         * 
599         * @throws HL7Exception
600         *             if the data fields in the message do not permit encoding
601         *             (e.g. required fields are null)
602         * @throws EncodingNotSupportedException
603         *             if the requested encoding is not supported by this parser.
604         */
605        protected String doEncode(Message source, String encoding) throws HL7Exception {
606                if (!this.supportsEncoding(encoding))
607                        throw new EncodingNotSupportedException("This parser does not support the " + encoding + " encoding");
608
609                return encode(source);
610        }
611
612        /**
613         * Formats a Message object into an HL7 message string using this parser's
614         * default encoding ("VB").
615         * 
616         * @throws HL7Exception
617         *             if the data fields in the message do not permit encoding
618         *             (e.g. required fields are null)
619         */
620        protected String doEncode(Message source) throws HL7Exception {
621                // get encoding characters ...
622                Segment msh = (Segment) source.get("MSH");
623                String fieldSepString = Terser.get(msh, 1, 0, 1, 1);
624
625                if (fieldSepString == null)
626                        throw new HL7Exception("Can't encode message: MSH-1 (field separator) is missing");
627
628                char fieldSep = '|';
629                if (fieldSepString.length() > 0) fieldSep = fieldSepString.charAt(0);
630
631                EncodingCharacters en = getValidEncodingCharacters(fieldSep, msh);
632
633                // pass down to group encoding method which will operate recursively on
634                // children ...
635                return encode(source, en, getParserConfiguration(), "");
636        }
637
638        private EncodingCharacters getValidEncodingCharacters(char fieldSep, Segment msh) throws HL7Exception {
639                String encCharString = Terser.get(msh, 2, 0, 1, 1);
640
641                if (encCharString == null) {
642                        throw new HL7Exception("Can't encode message: MSH-2 (encoding characters) is missing");
643                }
644
645                if (Version.V27.isGreaterThan(Version.versionOf(msh.getMessage().getVersion())) && encCharString.length() != 4) {
646                        throw new HL7Exception("Encoding characters (MSH-2) value '" + encCharString + "' invalid -- must be 4 characters", ErrorCode.DATA_TYPE_ERROR);
647                } else if (encCharString.length() != 4 && encCharString.length() != 5) {
648                        throw new HL7Exception("Encoding characters (MSH-2) value '" + encCharString + "' invalid -- must be 4 or 5 characters", ErrorCode.DATA_TYPE_ERROR);
649                }
650
651                return new EncodingCharacters(fieldSep, encCharString);
652        }
653
654        /**
655         * Returns given group serialized as a pipe-encoded string - this method is
656         * called by encode(Message source, String encoding).
657     *
658     * @param source group to be encoded
659     * @param encodingChars encoding characters to be used
660     * @throws HL7Exception if an error occurred while encoding
661     * @return encoded group
662         */
663        public static String encode(Group source, EncodingCharacters encodingChars) throws HL7Exception {
664                return encode(source, encodingChars, source.getMessage().getParser().getParserConfiguration(), "");
665        }
666
667        /**
668         * Returns given group serialized as a pipe-encoded string - this method is
669         * called by encode(Message source, String encoding).
670         */
671        private static String encode(Group source, EncodingCharacters encodingChars, ParserConfiguration parserConfiguration, String currentTerserPath) throws HL7Exception {
672                StringBuilder result = new StringBuilder();
673
674                String[] names = source.getNames();
675
676                String firstMandatorySegmentName = null;
677                boolean haveEncounteredMandatorySegment = false;
678                boolean haveEncounteredContent = false;
679                boolean haveHadMandatorySegment = false;
680                boolean haveHadSegmentBeforeMandatorySegment = false;
681
682                for (String nextName : names) {
683
684                        // source.get(nextName, 0);
685                        Structure[] reps = source.getAll(nextName);
686                        boolean nextNameIsRequired = source.isRequired(nextName);
687
688                        boolean havePreviouslyEncounteredMandatorySegment = haveEncounteredMandatorySegment;
689                        haveEncounteredMandatorySegment |= nextNameIsRequired;
690                        if (nextNameIsRequired && !haveHadMandatorySegment) {
691                                if (!source.isGroup(nextName)) {
692                                        firstMandatorySegmentName = nextName;
693                                }
694                        }
695
696                        String nextTerserPath = currentTerserPath.length() > 0 ? currentTerserPath + "/" + nextName : nextName;
697
698                        // Add all reps of the next segment/group
699                        for (Structure rep : reps) {
700
701                                if (rep instanceof Group) {
702
703                                        String encodedGroup = encode((Group) rep, encodingChars, parserConfiguration, nextTerserPath);
704                                        result.append(encodedGroup);
705
706                                        if (encodedGroup.length() > 0) {
707                                                if (!haveHadMandatorySegment && !haveEncounteredMandatorySegment) {
708                                                        haveHadSegmentBeforeMandatorySegment = true;
709                                                }
710                                                if (nextNameIsRequired && !haveHadMandatorySegment && !havePreviouslyEncounteredMandatorySegment) {
711                                                        haveHadMandatorySegment = true;
712                                                }
713                                                haveEncounteredContent = true;
714                                        }
715
716                                } else {
717
718                                        // Check if we are configured to force the encoding of this
719                                        // segment
720                                        boolean encodeEmptySegments = parserConfiguration.determineForcedEncodeIncludesTerserPath(nextTerserPath);
721                                        String segString = encode((Segment) rep, encodingChars, parserConfiguration, nextTerserPath);
722                                        if (segString.length() >= 4 || encodeEmptySegments) {
723                                                result.append(segString);
724
725                                                if (segString.length() == 3) {
726                                                        result.append(encodingChars.getFieldSeparator());
727                                                }
728
729                                                result.append(SEGMENT_DELIMITER);
730
731                                                haveEncounteredContent = true;
732
733                                                if (nextNameIsRequired) {
734                                                        haveHadMandatorySegment = true;
735                                                }
736
737                                                if (!haveHadMandatorySegment && !haveEncounteredMandatorySegment) {
738                                                        haveHadSegmentBeforeMandatorySegment = true;
739                                                }
740
741                                        }
742
743                                }
744
745                        }
746
747                }
748
749                if (firstMandatorySegmentName != null && !haveHadMandatorySegment && !haveHadSegmentBeforeMandatorySegment && haveEncounteredContent && parserConfiguration.isEncodeEmptyMandatorySegments()) {
750                        return firstMandatorySegmentName.substring(0, 3) + encodingChars.getFieldSeparator() + SEGMENT_DELIMITER + result;
751                } else {
752                        return result.toString();
753                }
754        }
755
756        /**
757         * Convenience factory method which returns an instance that has a new
758         * {@link DefaultHapiContext} initialized with a {@link NoValidation
759         * NoValidation validation context}.
760     *
761     * @return PipeParser with disabled validation
762         */
763        public static PipeParser getInstanceWithNoValidation() {
764                HapiContext context = new DefaultHapiContext();
765                context.setValidationContext(ValidationContextFactory.noValidation());
766                return new PipeParser(context);
767        }
768
769    /**
770     * Returns given segment serialized as a pipe-encoded string.
771     *
772     * @param source segment to be encoded
773     * @param encodingChars encoding characters to be used
774     * @return encoded group
775     */
776        public static String encode(Segment source, EncodingCharacters encodingChars) {
777                return encode(source, encodingChars, source.getMessage().getParser().getParserConfiguration(), null);
778        }
779
780        private static String encode(Segment source, EncodingCharacters encodingChars, ParserConfiguration parserConfig, String currentTerserPath) {
781                StringBuilder result = new StringBuilder();
782                result.append(source.getName());
783                result.append(encodingChars.getFieldSeparator());
784
785                // start at field 2 for MSH segment because field 1 is the field
786                // delimiter
787                int startAt = 1;
788                if (isDelimDefSegment(source.getName()))
789                        startAt = 2;
790
791                // loop through fields; for every field delimit any repetitions and add
792                // field delimiter after ...
793                int numFields = source.numFields();
794
795                int forceUpToFieldNum = 0;
796                if (parserConfig != null && currentTerserPath != null) {
797                        forceUpToFieldNum = parserConfig.determineForcedFieldNumForTerserPath(currentTerserPath);
798                }
799                numFields = Math.max(numFields, forceUpToFieldNum);
800
801                for (int i = startAt; i <= numFields; i++) {
802
803                        String nextFieldTerserPath = currentTerserPath + "-" + i;
804                        if (parserConfig != null && currentTerserPath != null) {
805                                for (String nextPath : parserConfig.getForcedEncode()) {
806                                        if (nextPath.startsWith(nextFieldTerserPath + "-")) {
807                                                try {
808                                                        source.getField(i, 0);
809                                                } catch (HL7Exception e) {
810                                                        log.error("Error while encoding segment: ", e);
811                                                }
812                                        }
813                                }
814                        }
815
816                        try {
817                                Type[] reps = source.getField(i);
818                                for (int j = 0; j < reps.length; j++) {
819                                        String fieldText = encode(reps[j], encodingChars, parserConfig, nextFieldTerserPath);
820                                        // if this is MSH-2, then it shouldn't be escaped, so
821                                        // unescape it again
822                                        if (isDelimDefSegment(source.getName()) && i == 2)
823                                                fieldText = parserConfig.getEscaping().unescape(fieldText, encodingChars);
824                                        result.append(fieldText);
825                                        if (j < reps.length - 1)
826                                                result.append(encodingChars.getRepetitionSeparator());
827                                }
828                        } catch (HL7Exception e) {
829                                log.error("Error while encoding segment: ", e);
830                        }
831                        result.append(encodingChars.getFieldSeparator());
832                }
833
834                // strip trailing delimiters ...
835                char fieldSeparator = encodingChars.getFieldSeparator();
836                String retVal = stripExtraDelimiters(result.toString(), fieldSeparator);
837
838                int offset = isDelimDefSegment(source.getName()) ? 1 : 0;
839                while (forceUpToFieldNum > 0 && (countInstancesOf(retVal, fieldSeparator) + offset) < forceUpToFieldNum) {
840                        retVal = retVal + fieldSeparator;
841                }
842
843                return retVal;
844        }
845
846        private static int countInstancesOf(String theString, char theCharToSearchFor) {
847                int retVal = 0;
848                for (int i = 0; i < theString.length(); i++) {
849                        if (theString.charAt(i) == theCharToSearchFor) {
850                                retVal++;
851                        }
852                }
853                return retVal;
854        }
855
856        /**
857         * Removes leading whitespace from the given string. This method was created
858         * to deal with frequent problems parsing messages that have been
859         * hand-written in windows. The intuitive way to delimit segments is to hit
860         * <ENTER> at the end of each segment, but this creates both a carriage
861         * return and a line feed, so to the parser, the first character of the next
862         * segment is the line feed.
863     *
864     * @param in input string
865     * @return string with leading whitespaces removed
866         */
867        public static String stripLeadingWhitespace(String in) {
868                StringBuilder out = new StringBuilder();
869                char[] chars = in.toCharArray();
870                int c = 0;
871                while (c < chars.length) {
872                        if (!Character.isWhitespace(chars[c]))
873                                break;
874                        c++;
875                }
876                for (int i = c; i < chars.length; i++) {
877                        out.append(chars[i]);
878                }
879                return out.toString();
880        }
881
882        /**
883         * <p>
884         * Returns a minimal amount of data from a message string, including only
885         * the data needed to send a response to the remote system. This includes
886         * the following fields:
887         * <ul>
888         * <li>field separator</li>
889         * <li>encoding characters</li>
890         * <li>processing ID</li>
891         * <li>message control ID</li>
892         * </ul>
893         * This method is intended for use when there is an error parsing a message,
894         * (so the Message object is unavailable) but an error message must be sent
895         * back to the remote system including some of the information in the
896         * inbound message. This method parses only that required information,
897         * hopefully avoiding the condition that caused the original error. The
898         * other fields in the returned MSH segment are empty.
899         * </p>
900         */
901        public Segment getCriticalResponseData(String message) throws HL7Exception {
902                // try to get MSH segment
903                int locStartMSH = message.indexOf("MSH");
904                if (locStartMSH < 0)
905                        throw new HL7Exception("Couldn't find MSH segment in message: " + message, ErrorCode.SEGMENT_SEQUENCE_ERROR);
906                int locEndMSH = message.indexOf('\r', locStartMSH + 1);
907                if (locEndMSH < 0)
908                        locEndMSH = message.length();
909                String mshString = message.substring(locStartMSH, locEndMSH);
910
911                // find out what the field separator is
912                char fieldSep = mshString.charAt(3);
913
914                // get field array
915                String[] fields = split(mshString, String.valueOf(fieldSep));
916
917                try {
918                        // parse required fields
919                        String encChars = fields[1];
920                        char compSep = encChars.charAt(0);
921                        String messControlID = fields[9];
922                        String[] procIDComps = split(fields[10], String.valueOf(compSep));
923
924                        // fill MSH segment
925                        String version = null;
926                        try {
927                                version = getVersion(message);
928                        } catch (Exception e) { /* use the default */
929                        }
930
931                        if (version == null) {
932                                Version availableVersion = Version.highestAvailableVersionOrDefault();
933                                version = availableVersion.getVersion();
934                        }
935
936                        Segment msh = Parser.makeControlMSH(version, getFactory());
937                        Terser.set(msh, 1, 0, 1, 1, String.valueOf(fieldSep));
938                        Terser.set(msh, 2, 0, 1, 1, encChars);
939                        Terser.set(msh, 10, 0, 1, 1, messControlID);
940                        Terser.set(msh, 11, 0, 1, 1, procIDComps[0]);
941                        Terser.set(msh, 12, 0, 1, 1, version);
942                        return msh;
943
944                } catch (Exception e) {
945                        throw new HL7Exception("Can't parse critical fields from MSH segment (" + e.getClass().getName() + ": " + e.getMessage() + "): " + mshString, ErrorCode.REQUIRED_FIELD_MISSING, e);
946                }
947
948        }
949
950        /**
951         * For response messages, returns the value of MSA-2 (the message ID of the
952         * message sent by the sending system). This value may be needed prior to
953         * main message parsing, so that (particularly in a multi-threaded scenario)
954         * the message can be routed to the thread that sent the request. We need
955         * this information first so that any parse exceptions are thrown to the
956         * correct thread. Returns null if MSA-2 can not be found (e.g. if the
957         * message is not a response message).
958         */
959        public String getAckID(String message) {
960                String ackID = null;
961                int startMSA = message.indexOf("\rMSA");
962                if (startMSA >= 0) {
963                        int startFieldOne = startMSA + 5;
964                        char fieldDelim = message.charAt(startFieldOne - 1);
965                        int start = message.indexOf(fieldDelim, startFieldOne) + 1;
966                        int end = message.indexOf(fieldDelim, start);
967                        int segEnd = message.indexOf(SEGMENT_DELIMITER, start);
968                        if (segEnd > start && segEnd < end)
969                                end = segEnd;
970
971                        // if there is no field delim after MSH-2, need to go to end of
972                        // message, but not including end seg delim if it exists
973                        if (end < 0) {
974                                if (message.charAt(message.length() - 1) == '\r') {
975                                        end = message.length() - 1;
976                                } else {
977                                        end = message.length();
978                                }
979                        }
980                        if (start > 0 && end > start) {
981                                ackID = message.substring(start, end);
982                        }
983                }
984                log.trace("ACK ID: {}", ackID);
985                return ackID;
986        }
987
988        /**
989         * Defaults to <code>false</code>
990         * 
991         * @see #isLegacyMode()
992         * @deprecated This will be removed in HAPI 3.0
993         */
994        public void setLegacyMode(boolean legacyMode) {
995                this.myLegacyMode = legacyMode;
996        }
997
998        /**
999         * {@inheritDoc }
1000         */
1001        @Override
1002        public String encode(Message source) throws HL7Exception {
1003                if (myLegacyMode != null && myLegacyMode) {
1004
1005                        @SuppressWarnings("deprecation")
1006                        OldPipeParser oldPipeParser = new OldPipeParser(getFactory());
1007
1008                        return oldPipeParser.encode(source);
1009                }
1010                return super.encode(source);
1011        }
1012
1013        /**
1014         * {@inheritDoc }
1015         */
1016        @Override
1017        public Message parse(String message) throws HL7Exception {
1018                if (myLegacyMode != null && myLegacyMode) {
1019
1020                        @SuppressWarnings("deprecation")
1021                        OldPipeParser oldPipeParser = new OldPipeParser(getFactory());
1022
1023                        return oldPipeParser.parse(message);
1024                }
1025                return super.parse(message);
1026        }
1027
1028        /**
1029         * <p>
1030         * Returns <code>true</code> if legacy mode is on.
1031         * </p>
1032         * <p>
1033         * Prior to release 1.0, when an unexpected segment was encountered in a
1034         * message, HAPI would recurse to the deepest nesting in the last group it
1035         * encountered after the current position in the message, and deposit the
1036         * segment there. This could lead to unusual behaviour where all segments
1037         * afterward would not be in an expected spot within the message.
1038         * </p>
1039         * <p>
1040         * This should normally be set to false, but any code written before the
1041         * release of HAPI 1.0 which depended on this behaviour might need legacy
1042         * mode to be set to true.
1043         * </p>
1044         * <p>
1045         * Defaults to <code>false</code>. Note that this method only overrides
1046         * behaviour of the {@link #parse(java.lang.String)} and
1047         * {@link #encode(ca.uhn.hl7v2.model.Message) } methods
1048         * </p>
1049         * 
1050         * @deprecated This will be removed in HAPI 3.0
1051         */
1052        public boolean isLegacyMode() {
1053                if (myLegacyMode == null)
1054                        return (Boolean.parseBoolean(System.getProperty(DEFAULT_LEGACY_MODE_PROPERTY)));
1055                return this.myLegacyMode;
1056        }
1057
1058        /**
1059         * Returns the version ID (MSH-12) from the given message, without fully
1060         * parsing the message. The version is needed prior to parsing in order to
1061         * determine the message class into which the text of the message should be
1062         * parsed.
1063         * 
1064         * @throws HL7Exception
1065         *             if the version field can not be found.
1066         */
1067        public String getVersion(String message) throws HL7Exception {
1068                int startMSH = message.indexOf("MSH");
1069                int endMSH = message.indexOf(PipeParser.SEGMENT_DELIMITER, startMSH);
1070                if (endMSH < 0)
1071                        endMSH = message.length();
1072                String msh = message.substring(startMSH, endMSH);
1073                String fieldSep;
1074                if (msh.length() > 3) {
1075                        fieldSep = String.valueOf(msh.charAt(3));
1076                } else {
1077                        throw new HL7Exception("Can't find field separator in MSH: " + msh, ErrorCode.UNSUPPORTED_VERSION_ID);
1078                }
1079
1080                String[] fields = split(msh, fieldSep);
1081
1082                String compSep;
1083                if (fields.length >= 2 && fields[1] != null && (fields[1].length() == 4 || fields[1].length() == 5)) {
1084                        compSep = String.valueOf(fields[1].charAt(0)); // get component separator as 1st encoding char
1085                } else {
1086                        throw new HL7Exception("Invalid or incomplete encoding characters - MSH-2 is " + fields[1], ErrorCode.REQUIRED_FIELD_MISSING);
1087                }
1088
1089                String version;
1090                if (fields.length >= 12) {
1091                        String[] comp = split(fields[11], compSep);
1092                        if (comp.length >= 1) {
1093                                version = comp[0];
1094                        } else {
1095                                throw new HL7Exception("Can't find version ID - MSH.12 is " + fields[11], ErrorCode.REQUIRED_FIELD_MISSING);
1096                        }
1097                } else if (getParserConfiguration().isAllowUnknownVersions()) {
1098                        return Version.highestAvailableVersionOrDefault().getVersion();
1099                } else {
1100                        throw new HL7Exception("Can't find version ID - MSH has only " + fields.length + " fields.", ErrorCode.REQUIRED_FIELD_MISSING);
1101                }
1102                return version;
1103        }
1104
1105        @Override
1106        public void parse(Message message, String string) throws HL7Exception {
1107                if (message instanceof AbstractSuperMessage && message.getName() == null) {
1108                        String structure = getStructure(string).messageStructure;
1109                        ((AbstractSuperMessage) message).setName(structure);
1110                }
1111
1112                message.setParser(this);
1113
1114                IStructureDefinition structureDef = getStructureDefinition(message);
1115                MessageIterator messageIter = new MessageIterator(message, structureDef, "MSH", true);
1116
1117                String[] segments = split(string, SEGMENT_DELIMITER);
1118
1119                if (segments.length == 0) {
1120                        throw new HL7Exception("Invalid message content: \"" + string + "\"");
1121                }
1122
1123                if (segments[0] == null || segments[0].length() < 4) {
1124                        throw new HL7Exception("Invalid message content: \"" + string + "\"");
1125                }
1126
1127                char delim = '|';
1128                String prevName = null;
1129                int repNum = 1;
1130                for (int i = 0; i < segments.length; i++) {
1131
1132                        // get rid of any leading whitespace characters ...
1133                        if (segments[i] != null && segments[i].length() > 0 && Character.isWhitespace(segments[i].charAt(0)))
1134                                segments[i] = stripLeadingWhitespace(segments[i]);
1135
1136                        // sometimes people put extra segment delimiters at end of msg ...
1137                        if (segments[i] != null && segments[i].length() >= 3) {
1138
1139                                final String name;
1140                                if (i == 0) {
1141                                        if (segments[i].length() < 4) {
1142                                                throw new HL7Exception("Invalid message content: \"" + string + "\"");
1143                                        }
1144                                        name = segments[i].substring(0, 3);
1145                                        delim = segments[i].charAt(3);
1146                                } else {
1147                                        if (segments[i].indexOf(delim) >= 0) {
1148                                                name = segments[i].substring(0, segments[i].indexOf(delim));
1149                                        } else {
1150                                                name = segments[i];
1151                                        }
1152                                }
1153
1154                                log.trace("Parsing segment {}", name);
1155
1156                                if (name.equals(prevName)) {
1157                                        repNum++;
1158                                } else {
1159                                        repNum = 1;
1160                                        prevName = name;
1161                                }
1162
1163                                messageIter.setDirection(name);
1164
1165                                try {
1166                                        if (messageIter.hasNext()) {
1167                                                Segment next = (Segment) messageIter.next();
1168                                                parse(next, segments[i], getEncodingChars(string), repNum);
1169                                        }
1170                                } catch (Error e) {
1171                                        if (e.getCause() instanceof HL7Exception) {
1172                                                throw (HL7Exception)e.getCause();
1173                                        }
1174                                        throw e;
1175                                }
1176                        }
1177                }
1178                
1179                applySuperStructureName(message);
1180        }
1181
1182        /**
1183         * A struct for holding a message class string and a boolean indicating
1184         * whether it was defined explicitly.
1185         */
1186        private static class MessageStructure {
1187                public final String messageStructure;
1188                public final boolean explicitlyDefined;
1189
1190                public MessageStructure(String theMessageStructure, boolean isExplicitlyDefined) {
1191                        messageStructure = theMessageStructure;
1192                        explicitlyDefined = isExplicitlyDefined;
1193                }
1194        }
1195
1196        private static class Holder<T> {
1197                private T myObject;
1198
1199                public T getObject() {
1200                        return myObject;
1201                }
1202
1203                public void setObject(T theObject) {
1204                        myObject = theObject;
1205                }
1206        }
1207
1208}