Merge pull request #306 from korpling/fix/exmaralda-tlis-without-times

Fix/exmaralda tlis without times

CHANGELOG.md

@@ -15,6 +15,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - `exmaralda` import now ranks order of tlis higher than sorting by time value (more compatible with modern EXMARaLDA files)
 
+### Fixed
+
+- `exmaralda` import keeps events with missing time values
+
 ## [0.15.0] - 2024-08-14
 
 ## [0.15.0] - 2024-08-14

docs/README.md

@@ -1,5 +1,5 @@
 | Type             | Modules                                                                                                                                                                                                                                                                                                                                                                                                                                                                     |
 |------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
 | Import formats   | [conllu](importers/conllu.md), [exmaralda](importers/exmaralda.md), [graphml](importers/graphml.md), [meta](importers/meta.md), [none](importers/none.md), [opus](importers/opus.md), [path](importers/path.md), [ptb](importers/ptb.md), [relannis](importers/relannis.md), [saltxml](importers/saltxml.md), [textgrid](importers/textgrid.md), [toolbox](importers/toolbox.md), [treetagger](importers/treetagger.md), [xlsx](importers/xlsx.md), [xml](importers/xml.md) |
-| Export formats   | [graphml](exporters/graphml.md), [exmaralda](exporters/exmaralda.md), [sequence](exporters/sequence.md), [table](exporters/table.md), [textgrid](exporters/textgrid.md), [xlsx](exporters/xlsx.md)                                                                                                                                                                                                                                                                          |
+| Export formats   | [conllu](exporters/conllu.md), [graphml](exporters/graphml.md), [exmaralda](exporters/exmaralda.md), [sequence](exporters/sequence.md), [table](exporters/table.md), [textgrid](exporters/textgrid.md), [xlsx](exporters/xlsx.md)                                                                                                                                                                                                                                           |
 | Graph operations | [check](graph_ops/check.md), [collapse](graph_ops/collapse.md), [filter](graph_ops/filter.md), [visualize](graph_ops/visualize.md), [enumerate](graph_ops/enumerate.md), [link](graph_ops/link.md), [map](graph_ops/map.md), [revise](graph_ops/revise.md), [chunk](graph_ops/chunk.md), [split](graph_ops/split.md), [none](graph_ops/none.md)                                                                                                                             |

docs/exporters/conllu.md

@@ -0,0 +1,151 @@
+# conllu (exporter)
+
+This module exports a graph in CoNLL-U format.
+
+## Configuration
+
+###  doc
+
+This key is used to determine nodes that whose part-of subgraph constitutes a document, i. e. the entire input for a file.
+Default is `annis::doc`, or `{ ns = "annis", name = "doc" }`.
+
+Example:
+```toml
+[export.config]
+doc = "annis::doc"
+```
+
+###  groupby
+
+This optional annotation key is used to identify annotation spans, that constitute a sentence. Default is no export of sentence blocks.
+Default is `annis::doc`, or `{ ns = "annis", name = "doc" }`.
+
+Example:
+```toml
+[export.config]
+groupby = "norm::sentence"
+```
+
+###  ordering
+
+The nodes connected by this annotation component are used as nodes defining a line in a CoNLL-U file. Usually you want to use an ordering.
+Default is `{ ctype = "Ordering", layer = "annis", name = "" }`.
+
+Example:
+```toml
+[export.config]
+ordering = { ctype = "Ordering", layer = "annis", name = "norm" }
+```
+
+###  form
+
+This annotation key is used to write the form column.
+Default is `{ ns = "annis", name = "tok" }`.
+
+Example:
+```toml
+[export.config]
+form = { ns = "norm", name = "norm" }
+```
+
+###  lemma
+
+This annotation key is used to write the lemma column.
+Default is `{ ns = "", name = "tok" }`.
+
+Example:
+```toml
+[export.config]
+lemma = { ns = "norm", name = "lemma" }
+```
+
+###  upos
+
+This annotation key is used to write the upos column.
+Default is `{ ns = "", name = "upos" }`.
+
+Example:
+```toml
+[export.config]
+upos = { ns = "norm", name = "pos" }
+```
+
+###  xpos
+
+This annotation key is used to write the xpos column.
+Default is `{ ns = "", name = "xpos" }`.
+
+Example:
+```toml
+[export.config]
+upos = { ns = "norm", name = "pos_spec" }
+```
+
+###  features
+
+This list of annotation keys will be represented in the feature column.
+Default is the empty list.
+
+Example:
+```toml
+[export.config]
+features = ["Animacy", "Tense", "VerbClass"]
+```
+
+###  dependency_component
+
+The nodes connected by this annotation component are used to export dependencies.
+Default is none, so nothing will be exported.
+
+Example:
+```toml
+[export.config]
+dependency_component = { ctype = "Pointing", layer = "", name = "dependencies" }
+```
+
+###  dependency_anno
+
+This annotation key is used to write the dependency relation, which will be looked for on the dependency edges.
+Default is none, so nothing will be exported.
+
+Example:
+```toml
+[export.config]
+dependency_anno = { ns = "", name = "deprel" }
+```
+
+###  enhanced_components
+
+The listed components will be used to export enhanced dependencies. More than
+one component can be listed.
+Default is the empty list, so nothing will be exported.
+
+Example:
+```toml
+[export.config]
+enhanced_components = [{ ctype = "Pointing", layer = "", name = "dependencies" }]
+```
+
+###  enhanced_annos
+
+This list of annotation keys defines the annotation keys, that correspond to the
+edge labels in the component listed in `enhanced_components`. The i-th element of
+one list belongs to the i-th element in the other list. Default is the empty list.
+
+Example:
+```toml
+[export.config]
+enhanced_annos = ["func"]
+```
+
+###  misc
+
+This list of annotation keys will be represented in the misc column.
+Default is the empty list.
+
+Example:
+```toml
+[export.config]
+misc = ["NoSpaceAfter", "Referent"]
+```
+

docs/exporters/table.md

@@ -39,6 +39,16 @@ Example:
 quote_char = "\""
 ```
 
+###  no_value
+
+Provides the string sequence used for n/a. Default is the empty string.
+
+Example:
+```toml
+[export.config]
+no_value = "n/a"
+```
+
 ###  ingoing
 
 By listing annotation components, the ingoing edges of that component and their annotations

src/importer/exmaralda/mod.rs

@@ -93,8 +93,7 @@ impl ImportEXMARaLDA {
         let mut speaker_map = BTreeMap::new();
         let mut parent_map: BTreeMap<String, BTreeMap<String, String>> = BTreeMap::new();
         let mut already_defined: BTreeSet<String> = BTreeSet::new();
-        let mut named_orderings: BTreeMap<String, Vec<(OrderedFloat<f64>, String)>> =
-            BTreeMap::new();
+        let mut named_orderings: BTreeMap<String, Vec<(usize, String)>> = BTreeMap::new();
         let mut tlis = Vec::new();
         // reader
         let f = File::open(document_path)?;
@@ -441,18 +440,6 @@ impl ImportEXMARaLDA {
                                 "{}#{}_{}_{}-{}",
                                 doc_node_name, tier_type, speaker_id, start_id, end_id
                             ); // this is not a unique id as not intended to be
-                            let start_time = if let Some((Some(t), _)) = timeline.get(key) {
-                                t
-                            } else {
-                                if let Some(sender) = tx {
-                                    let msg = format!(
-                                            "Could not determine start time of event {}::{}:{}-{}. Event will be skipped.",
-                                            &speaker_id, &anno_name, &start_id, &end_id
-                                        );
-                                    sender.send(StatusMessage::Warning(msg))?;
-                                }
-                                continue;
-                            };
                             if !already_defined.contains(&node_name) {
                                 update.add_event(UpdateEvent::AddNode {
                                     node_name: node_name.to_string(),
@@ -492,7 +479,9 @@ impl ImportEXMARaLDA {
                                     }
                                     continue;
                                 };
-                                if let Some((Some(end_time), _)) = node_tpl {
+                                if let (Some((Some(start_time), _)), Some((Some(end_time), _))) =
+                                    (timeline.get(key), node_tpl)
+                                {
                                     update.add_event(UpdateEvent::AddNodeLabel {
                                         node_name: node_name.to_string(),
                                         anno_ns: ANNIS_NS.to_string(),
@@ -517,7 +506,7 @@ impl ImportEXMARaLDA {
                                     anno_value: text.to_string(),
                                 })?;
                                 // order nodes
-                                let order_tpl = (*start_time, node_name.to_string());
+                                let order_tpl = (start_i, node_name.to_string());
                                 match named_orderings.entry(anno_name.to_string()) {
                                     std::collections::btree_map::Entry::Vacant(e) => {
                                         e.insert(vec![order_tpl]);
@@ -574,10 +563,7 @@ impl ImportEXMARaLDA {
         // build order relations
         for (name, node_name_vec) in named_orderings {
             let mut prev = None;
-            for (_, node_name) in node_name_vec
-                .into_iter()
-                .sorted_by(|a, b| a.0.total_cmp(&b.0))
-            {
+            for (_, node_name) in node_name_vec.into_iter().sorted_by(|a, b| a.0.cmp(&b.0)) {
                 if let Some(source) = prev {
                     update.add_event(UpdateEvent::AddEdge {
                         source_node: source,