Skip to main content

edifact_mapper/
mapper.rs

1//! High-level [`Mapper`] API for EDIFACT-to-BO4E conversion.
2
3use std::collections::HashMap;
4use std::sync::Mutex;
5
6use mig_assembly::ConversionService;
7use mig_bo4e::engine::DataBundle;
8use mig_bo4e::MappingEngine;
9
10use crate::data_dir::DataDir;
11use crate::error::MapperError;
12
13/// Result of a BO4E mapping operation.
14pub struct Bo4eResult {
15    /// The PID (Pruefidentifikator) that was detected or specified.
16    pub pid: String,
17    /// The EDIFACT message type (e.g., "UTILMD", "MSCONS").
18    pub message_type: String,
19    /// The message variant (e.g., "UTILMD_Strom", "MSCONS").
20    pub variant: String,
21    /// The mapped BO4E JSON output.
22    pub bo4e: serde_json::Value,
23}
24
25/// High-level facade for EDIFACT-to-BO4E conversion.
26///
27/// Wraps [`DataBundle`] loading with lazy/eager initialization, and provides
28/// convenient accessors for [`ConversionService`] and [`MappingEngine`] instances.
29///
30/// # Example
31///
32/// ```ignore
33/// use edifact_mapper::{DataDir, Mapper};
34///
35/// let mapper = Mapper::from_data_dir(
36///     DataDir::path("/opt/edifact/data").eager(&["FV2504"])
37/// )?;
38///
39/// let cs = mapper.conversion_service("FV2504", "UTILMD_Strom")?;
40/// let engine = mapper.engine("FV2504", "UTILMD_Strom", "55001")?;
41/// ```
42pub struct Mapper {
43    data_dir: DataDir,
44    bundles: Mutex<HashMap<String, DataBundle>>,
45}
46
47impl Mapper {
48    /// Create a new `Mapper` from a [`DataDir`] configuration.
49    ///
50    /// Any format versions marked as [`eager`](DataDir::eager) are loaded immediately.
51    /// All others are loaded lazily on first access.
52    pub fn from_data_dir(data_dir: DataDir) -> Result<Self, MapperError> {
53        let mapper = Self {
54            data_dir,
55            bundles: Mutex::new(HashMap::new()),
56        };
57        let eager_fvs: Vec<String> = mapper.data_dir.eager_fvs().to_vec();
58        for fv in &eager_fvs {
59            mapper.ensure_bundle_loaded(fv)?;
60        }
61        Ok(mapper)
62    }
63
64    /// Ensure that the bundle for `fv` is loaded into memory.
65    fn ensure_bundle_loaded(&self, fv: &str) -> Result<(), MapperError> {
66        let mut bundles = self.bundles.lock().unwrap();
67        if bundles.contains_key(fv) {
68            return Ok(());
69        }
70        let path = self.data_dir.bundle_path(fv);
71        if !path.exists() {
72            return Err(MapperError::BundleNotFound { fv: fv.to_string() });
73        }
74        let bundle = DataBundle::load(&path)?;
75        bundles.insert(fv.to_string(), bundle);
76        Ok(())
77    }
78
79    /// Get a [`ConversionService`] for the given format version and variant.
80    ///
81    /// The service can tokenize EDIFACT input and assemble it into a MIG tree.
82    pub fn conversion_service(
83        &self,
84        fv: &str,
85        variant: &str,
86    ) -> Result<ConversionService, MapperError> {
87        self.ensure_bundle_loaded(fv)?;
88        let bundles = self.bundles.lock().unwrap();
89        let bundle = bundles.get(fv).unwrap();
90        let vc = bundle
91            .variant(variant)
92            .ok_or_else(|| MapperError::VariantNotFound {
93                fv: fv.to_string(),
94                variant: variant.to_string(),
95            })?;
96        let mig = vc
97            .mig_schema
98            .as_ref()
99            .ok_or_else(|| MapperError::VariantNotFound {
100                fv: fv.to_string(),
101                variant: format!("{variant} (no MIG schema in bundle)"),
102            })?;
103        Ok(ConversionService::from_mig(mig.clone()))
104    }
105
106    /// Get a [`MappingEngine`] for a specific PID within a format version and variant.
107    ///
108    /// The engine can convert between assembled MIG trees and BO4E JSON.
109    pub fn engine(&self, fv: &str, variant: &str, pid: &str) -> Result<MappingEngine, MapperError> {
110        self.ensure_bundle_loaded(fv)?;
111        let bundles = self.bundles.lock().unwrap();
112        let bundle = bundles.get(fv).unwrap();
113        let vc = bundle
114            .variant(variant)
115            .ok_or_else(|| MapperError::VariantNotFound {
116                fv: fv.to_string(),
117                variant: variant.to_string(),
118            })?;
119        let pid_key = format!("pid_{pid}");
120        let defs = vc
121            .combined_defs
122            .get(&pid_key)
123            .ok_or_else(|| MapperError::PidNotFound {
124                fv: fv.to_string(),
125                variant: variant.to_string(),
126                pid: pid.to_string(),
127            })?;
128        Ok(MappingEngine::from_definitions(defs.clone()))
129    }
130
131    /// Validate a BO4E JSON object against PID requirements.
132    ///
133    /// Returns a list of validation errors. Empty list = valid.
134    /// The `json` should be the transaction-level stammdaten (the entity map).
135    pub fn validate_pid(
136        &self,
137        json: &serde_json::Value,
138        fv: &str,
139        variant: &str,
140        pid: &str,
141    ) -> Result<Vec<mig_bo4e::PidValidationError>, MapperError> {
142        self.ensure_bundle_loaded(fv)?;
143        let bundles = self.bundles.lock().unwrap();
144        let bundle = bundles.get(fv).unwrap();
145        let vc = bundle
146            .variant(variant)
147            .ok_or_else(|| MapperError::VariantNotFound {
148                fv: fv.to_string(),
149                variant: variant.to_string(),
150            })?;
151        let pid_key = format!("pid_{pid}");
152        let requirements =
153            vc.pid_requirements
154                .get(&pid_key)
155                .ok_or_else(|| MapperError::PidNotFound {
156                    fv: fv.to_string(),
157                    variant: variant.to_string(),
158                    pid: pid.to_string(),
159                })?;
160
161        Ok(mig_bo4e::pid_validation::validate_pid_json(
162            json,
163            requirements,
164        ))
165    }
166
167    /// Validate a typed BO4E struct against PID requirements.
168    ///
169    /// Convenience wrapper that serializes the struct to JSON first.
170    /// Works with any `Pid*Interchange` or `Pid*MessageStammdaten` type.
171    ///
172    /// # Example
173    /// ```ignore
174    /// let interchange = build_55001_interchange();
175    /// let errors = mapper.validate_pid_struct(&interchange, "FV2504", "UTILMD_Strom", "55001")?;
176    /// assert!(errors.is_empty(), "Errors:\n{}", ValidationReport(errors));
177    /// ```
178    pub fn validate_pid_struct(
179        &self,
180        value: &impl serde::Serialize,
181        fv: &str,
182        variant: &str,
183        pid: &str,
184    ) -> Result<Vec<mig_bo4e::PidValidationError>, MapperError> {
185        let json = serde_json::to_value(value).map_err(|e| {
186            MapperError::Mapping(mig_bo4e::MappingError::TypeConversion(e.to_string()))
187        })?;
188        self.validate_pid(&json, fv, variant, pid)
189    }
190
191    /// Validate with AHB condition awareness.
192    ///
193    /// Reverse-maps the JSON to EDIFACT segments, evaluates AHB conditions,
194    /// and reports fields as required/optional based on the actual data present.
195    ///
196    /// Falls back to basic validation (without conditions) if no condition
197    /// evaluator is available for the given variant/format version combination.
198    pub fn validate_pid_with_conditions(
199        &self,
200        json: &serde_json::Value,
201        fv: &str,
202        variant: &str,
203        pid: &str,
204    ) -> Result<Vec<mig_bo4e::PidValidationError>, MapperError> {
205        self.ensure_bundle_loaded(fv)?;
206        let bundles = self.bundles.lock().unwrap();
207        let bundle = bundles.get(fv).unwrap();
208        let vc = bundle
209            .variant(variant)
210            .ok_or_else(|| MapperError::VariantNotFound {
211                fv: fv.to_string(),
212                variant: variant.to_string(),
213            })?;
214        let pid_key = format!("pid_{pid}");
215
216        let requirements =
217            vc.pid_requirements
218                .get(&pid_key)
219                .ok_or_else(|| MapperError::PidNotFound {
220                    fv: fv.to_string(),
221                    variant: variant.to_string(),
222                    pid: pid.to_string(),
223                })?;
224
225        // Try to get a condition evaluator for this variant
226        let evaluator = crate::evaluator_factory::create_evaluator(variant, fv);
227
228        if let Some(evaluator) = evaluator {
229            // Reverse-map JSON to EDIFACT segments for condition evaluation context
230            let defs = vc
231                .combined_defs
232                .get(&pid_key)
233                .ok_or_else(|| MapperError::PidNotFound {
234                    fv: fv.to_string(),
235                    variant: variant.to_string(),
236                    pid: pid.to_string(),
237                })?;
238            let engine = MappingEngine::from_definitions(defs.clone());
239            let tree = engine.map_all_reverse(json, None);
240
241            // Convert AssembledTree to flat OwnedSegments for EvaluationContext
242            let segments = crate::tree_to_segments::tree_to_owned_segments(&tree);
243
244            // Validate with condition awareness
245            Ok(crate::evaluator_factory::validate_with_boxed_evaluator(
246                evaluator.as_ref(),
247                json,
248                requirements,
249                pid,
250                &segments,
251            ))
252        } else {
253            // No evaluator available — fall back to basic validation
254            Ok(mig_bo4e::pid_validation::validate_pid_json(
255                json,
256                requirements,
257            ))
258        }
259    }
260
261    /// Convert BO4E JSON back to an EDIFACT string.
262    ///
263    /// Takes message-level stammdaten, a slice of per-transaction stammdaten,
264    /// and produces an EDIFACT message body (UNH through UNT content segments,
265    /// without UNB/UNZ interchange envelope).
266    ///
267    /// # Arguments
268    ///
269    /// * `msg_stammdaten` — message-level entities (e.g., Marktteilnehmer from SG2)
270    /// * `tx_stammdaten` — per-transaction entities (one per transaction/SG4 instance)
271    /// * `fv` — format version (e.g., "FV2504")
272    /// * `variant` — message variant (e.g., "UTILMD_Strom")
273    /// * `pid` — Pruefidentifikator (e.g., "55001")
274    ///
275    /// # Example
276    ///
277    /// ```ignore
278    /// let edifact = mapper.to_edifact(
279    ///     &msg_json,
280    ///     &[tx_json],
281    ///     "FV2504",
282    ///     "UTILMD_Strom",
283    ///     "55001",
284    /// )?;
285    /// ```
286    pub fn to_edifact(
287        &self,
288        msg_stammdaten: &serde_json::Value,
289        tx_stammdaten: &[serde_json::Value],
290        fv: &str,
291        variant: &str,
292        pid: &str,
293    ) -> Result<String, MapperError> {
294        self.ensure_bundle_loaded(fv)?;
295        let bundles = self.bundles.lock().unwrap();
296        let bundle = bundles.get(fv).unwrap();
297        let vc = bundle
298            .variant(variant)
299            .ok_or_else(|| MapperError::VariantNotFound {
300                fv: fv.to_string(),
301                variant: variant.to_string(),
302            })?;
303
304        let tx_group = vc
305            .tx_group(pid)
306            .ok_or_else(|| MapperError::PidNotFound {
307                fv: fv.to_string(),
308                variant: variant.to_string(),
309                pid: pid.to_string(),
310            })?;
311
312        let msg_engine = vc.msg_engine();
313        let tx_engine =
314            vc.tx_engine(pid)
315                .ok_or_else(|| MapperError::PidNotFound {
316                    fv: fv.to_string(),
317                    variant: variant.to_string(),
318                    pid: pid.to_string(),
319                })?;
320
321        let filtered_mig =
322            vc.filtered_mig(pid)
323                .ok_or_else(|| MapperError::NoMigSchema {
324                    fv: fv.to_string(),
325                    variant: variant.to_string(),
326                })?;
327
328        // Build MappedMessage from the provided JSON
329        let transaktionen: Vec<mig_bo4e::model::MappedTransaktion> = tx_stammdaten
330            .iter()
331            .map(|tx| mig_bo4e::model::MappedTransaktion {
332                stammdaten: tx.clone(),
333                nesting_info: Default::default(),
334            })
335            .collect();
336        let mapped = mig_bo4e::model::MappedMessage {
337            stammdaten: msg_stammdaten.clone(),
338            transaktionen,
339            nesting_info: Default::default(),
340        };
341
342        // Reverse map → AssembledTree
343        let tree = MappingEngine::map_interchange_reverse(
344            &msg_engine,
345            &tx_engine,
346            &mapped,
347            &tx_group,
348            Some(&filtered_mig),
349        );
350
351        // Disassemble → ordered segments
352        let disassembler =
353            mig_assembly::disassembler::Disassembler::new(&filtered_mig);
354        let segments = disassembler.disassemble(&tree);
355
356        // Render to EDIFACT string with default delimiters
357        let delimiters = edifact_primitives::EdifactDelimiters::default();
358        Ok(mig_assembly::renderer::render_edifact(
359            &segments,
360            &delimiters,
361        ))
362    }
363
364    /// Convert a typed BO4E struct to an EDIFACT string.
365    ///
366    /// Convenience wrapper that serializes the struct to JSON first.
367    /// The struct should serialize to the `Nachricht` shape:
368    /// `{ "stammdaten": {...}, "transaktionen": [{...}] }`
369    pub fn to_edifact_struct(
370        &self,
371        nachricht: &impl serde::Serialize,
372        fv: &str,
373        variant: &str,
374        pid: &str,
375    ) -> Result<String, MapperError> {
376        let json = serde_json::to_value(nachricht)
377            .map_err(|e| MapperError::Serialization(e.to_string()))?;
378
379        let msg_stammdaten = json
380            .get("stammdaten")
381            .cloned()
382            .unwrap_or(serde_json::Value::Object(Default::default()));
383
384        let tx_stammdaten: Vec<serde_json::Value> = json
385            .get("transaktionen")
386            .and_then(|v| v.as_array())
387            .cloned()
388            .unwrap_or_default();
389
390        self.to_edifact(&msg_stammdaten, &tx_stammdaten, fv, variant, pid)
391    }
392
393    /// List all format versions currently loaded in memory.
394    pub fn loaded_format_versions(&self) -> Vec<String> {
395        self.bundles.lock().unwrap().keys().cloned().collect()
396    }
397
398    /// List all variants available in a format version's bundle.
399    ///
400    /// Loads the bundle if not already loaded.
401    pub fn variants(&self, fv: &str) -> Result<Vec<String>, MapperError> {
402        self.ensure_bundle_loaded(fv)?;
403        let bundles = self.bundles.lock().unwrap();
404        let bundle = bundles.get(fv).unwrap();
405        Ok(bundle.variants.keys().cloned().collect())
406    }
407}
408
409
410#[cfg(test)]
411mod tests {
412    use super::*;
413    use std::path::Path;
414
415    fn data_dir() -> Option<std::path::PathBuf> {
416        // Try dist/ first (pre-built data bundles), then cache/mappings/
417        let dist = Path::new(env!("CARGO_MANIFEST_DIR")).join("../../dist");
418        if dist.join("edifact-data-FV2504.bin").exists() {
419            return Some(dist);
420        }
421        let cache = Path::new(env!("CARGO_MANIFEST_DIR")).join("../../cache/mappings");
422        if cache.join("FV2504").exists() {
423            return Some(cache);
424        }
425        eprintln!("Skipping test: no DataBundle files found");
426        None
427    }
428
429    #[test]
430    fn test_to_edifact_produces_edifact_output() {
431        let Some(data_dir) = data_dir() else {
432            return;
433        };
434        let mapper =
435            Mapper::from_data_dir(DataDir::path(&data_dir).eager(&["FV2504"])).unwrap();
436
437        let msg_stammdaten = serde_json::json!({
438            "marktteilnehmer": [{
439                "marktrolle": "MS",
440                "rollencodenummer": "9900123456789",
441                "codepflegeCode": "293"
442            }]
443        });
444        let tx_stammdaten = serde_json::json!({
445            "prozessdaten": {
446                "pruefidentifikator": "55001",
447                "vorgangId": "ABC123",
448                "transaktionsgrund": "E01"
449            }
450        });
451
452        let result = mapper.to_edifact(
453            &msg_stammdaten,
454            &[tx_stammdaten],
455            "FV2504",
456            "UTILMD_Strom",
457            "55001",
458        );
459        assert!(result.is_ok(), "to_edifact failed: {:?}", result.err());
460        let edifact = result.unwrap();
461        assert!(!edifact.is_empty(), "EDIFACT output should not be empty");
462        // Should produce NAD segment from marktteilnehmer
463        assert!(edifact.contains("NAD"), "Should contain NAD segment");
464        // Should produce IDE segment from prozessdaten
465        assert!(edifact.contains("IDE"), "Should contain IDE segment");
466    }
467
468    #[test]
469    fn test_to_edifact_struct_produces_edifact_output() {
470        let Some(data_dir) = data_dir() else {
471            return;
472        };
473        let mapper =
474            Mapper::from_data_dir(DataDir::path(&data_dir).eager(&["FV2504"])).unwrap();
475
476        let nachricht = serde_json::json!({
477            "stammdaten": {
478                "marktteilnehmer": [{
479                    "marktrolle": "MS",
480                    "rollencodenummer": "9900123456789",
481                    "codepflegeCode": "293"
482                }]
483            },
484            "transaktionen": [{
485                "prozessdaten": {
486                    "pruefidentifikator": "55001",
487                    "vorgangId": "ABC123"
488                }
489            }]
490        });
491
492        let result = mapper.to_edifact_struct(&nachricht, "FV2504", "UTILMD_Strom", "55001");
493        assert!(
494            result.is_ok(),
495            "to_edifact_struct failed: {:?}",
496            result.err()
497        );
498        let edifact = result.unwrap();
499        assert!(!edifact.is_empty(), "EDIFACT output should not be empty");
500    }
501
502    #[test]
503    fn test_to_edifact_invalid_fv_returns_error() {
504        let Some(data_dir) = data_dir() else {
505            return;
506        };
507        let mapper =
508            Mapper::from_data_dir(DataDir::path(&data_dir).eager(&["FV2504"])).unwrap();
509
510        let result = mapper.to_edifact(
511            &serde_json::json!({}),
512            &[serde_json::json!({})],
513            "FV9999",
514            "UTILMD_Strom",
515            "55001",
516        );
517        assert!(result.is_err());
518    }
519
520    #[test]
521    fn test_to_edifact_invalid_variant_returns_error() {
522        let Some(data_dir) = data_dir() else {
523            return;
524        };
525        let mapper =
526            Mapper::from_data_dir(DataDir::path(&data_dir).eager(&["FV2504"])).unwrap();
527
528        let result = mapper.to_edifact(
529            &serde_json::json!({}),
530            &[serde_json::json!({})],
531            "FV2504",
532            "NONEXISTENT",
533            "55001",
534        );
535        assert!(result.is_err());
536    }
537
538    #[test]
539    fn test_to_edifact_invalid_pid_returns_error() {
540        let Some(data_dir) = data_dir() else {
541            return;
542        };
543        let mapper =
544            Mapper::from_data_dir(DataDir::path(&data_dir).eager(&["FV2504"])).unwrap();
545
546        let result = mapper.to_edifact(
547            &serde_json::json!({}),
548            &[serde_json::json!({})],
549            "FV2504",
550            "UTILMD_Strom",
551            "99999",
552        );
553        assert!(result.is_err());
554    }
555}