mailwoman 0.3.0 → 1.0.0

package/README.md

@@ -2,50 +2,33 @@
 
 A natural language classification engine for geocoding.
 
-![GitHub License](https://img.shields.io/github/license/isp-nexus/mailwoman)
-![GitHub commit activity](https://img.shields.io/github/commit-activity/m/isp-nexus/mailwoman)
+![GitHub License](https://img.shields.io/github/license/sister-software/mailwoman)
+![GitHub commit activity](https://img.shields.io/github/commit-activity/m/sister-software/mailwoman)
 
-# Pelias Parser
-
-## AddressParser Example
-
-```
-30 w 26 st nyc 10010
-
-(0.95) ➜ [
-  { housenumber: '30' },
-  { street: 'w 26 st' },
-  { locality: 'nyc' },
-  { postcode: '10010' }
+```js
+// npx mailwoman debug "Mt Tabor Park, 6220 SE Salmon St, Portland, OR 97215, USA"
+
+;[
+	{ venue: "Mt Tabor Park", confidence: 0.8, offset: 0, penalty: 0 },
+	{ house_number: "6220", confidence: 0.9, offset: 15, penalty: 0 },
+	{ street: "SE Salmon St", confidence: 0.98, offset: 20, penalty: 0 },
+	{ locality: "Portland", confidence: 1, offset: 34, penalty: 0 },
+	{ region: "OR", confidence: 1, offset: 44, penalty: 0 },
+	{ postcode: "97215", confidence: 1, offset: 47, penalty: 0 },
+	{ country: "USA", confidence: 0.9, offset: 54, penalty: 0 },
 ]
 ```
 
-## Application Interfaces
-
-You can access the library via three different interfaces:
-
-- all parts of the codebase are available in `javascript` via `npm`
-- on the `command line` via the `node bin/cli.js` script
-- through a `web service` via the `node server/http.js` script
-
-> the web service provides an interactive demo at the URL `/parser/parse`
-
 ## Quick Start
 
 A quick and easy way to get started with the library is to use the command-line interface:
 
 ```
-node bin/cli.js West 26th Street, New York, NYC, 10010
+npx mailwoman parse "West 26th Street, New York, NYC, 10010"
 ```
 
-![cli](./docs/cli.png)
-
----
-
 # Architecture Description
 
-Please refer to the CLI screenshot above for a visual reference.
-
 ## Tokenization
 
 Tokenization is the process of splitting text into individual words.
@@ -95,11 +78,11 @@ The graph is free-form, so it's easy to add a new relationship between terms in
 Graph Example:
 
 ```js
-// find the next word in this section
-word.findOne("next")
+// Find the next word in this section
+word.next
 
-// find all words in this phrase
-phrase.findAll("child")
+// Find all words in this phrase
+phrase.children
 ```
 
 ## Classification
@@ -133,24 +116,24 @@ You can find them in the `/classifier` directory, dictionary-based classifiers u
 Example of some of the included classifiers:
 
 - Word Classifiers
-  - HouseNumberClassifier
-  - PostcodeClassifier
-  - StreetPrefixClassifier
-  - StreetSuffixClassifier
-  - CompoundStreetClassifier
-  - DirectionalClassifier
-  - OrdinalClassifier
-  - StopWordClassifier
+  - `house_number`
+  - `postcode`
+  - `street_prefix`
+  - `street_suffix`
+  - `compound_street`
+  - `directional`
+  - `ordinal`
+  - `stop_word`
 - Phrase Classifiers
-  - IntersectionClassifier
-  - PersonClassifier
-  - GivenNameClassifier
-  - SurnameClassifier
-  - PersonalSuffixClassifier
-  - PersonalTitleClassifier
-  - ChainClassifier
-  - PlaceClassifier
-  - WhosOnFirstClassifier
+  - `intersection`
+  - `person`
+  - `given_name`
+  - `surname`
+  - `personal_suffix`
+  - `personal_title`
+  - `chain`
+  - `place`
+  - `whos_on_first`
 
 ## Solvers
 
@@ -201,8 +184,8 @@ Unmodified portions of Mailwoman derived from Pelias Parser remain under the
 MIT license.
 
 For more information on commercial usage licensing, please contact us at
-`hello@isp.nexus`
+`hello@sister.software`
 
 # Acknowledgements
 
-This project was built by the [Open ISP](https://isp.nexus) team and was made possible by the contributions of the Pelias community.
+This project was made possible by contributions of the Pelias community.

package/out/classifiers/AdjacencyClassifier.d.ts

@@ -0,0 +1,16 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { SectionClassifier, Span } from "mailwoman/core";
+/**
+ * An adjacency classifier is used to find three adjacent words in a section.
+ *
+ * The first word must be a house number, the second word must be a street, and the third word must
+ * be a street suffix.
+ */
+export declare class AdjacencyClassifier extends SectionClassifier {
+    explore(section: Span): void;
+}
+//# sourceMappingURL=AdjacencyClassifier.d.ts.map

package/out/classifiers/AdjacencyClassifier.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"AdjacencyClassifier.d.ts","sourceRoot":"","sources":["../../classifiers/AdjacencyClassifier.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,iBAAiB,EAAE,IAAI,EAAE,MAAM,gBAAgB,CAAA;AAExD;;;;;GAKG;AACH,qBAAa,mBAAoB,SAAQ,iBAAiB;IACzD,OAAO,CAAC,OAAO,EAAE,IAAI,GAAG,IAAI;CA6C5B"}

package/out/classifiers/AdjacencyClassifier.js

@@ -0,0 +1,50 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { SectionClassifier } from "mailwoman/core";
+/**
+ * An adjacency classifier is used to find three adjacent words in a section.
+ *
+ * The first word must be a house number, the second word must be a street, and the third word must
+ * be a street suffix.
+ */
+export class AdjacencyClassifier extends SectionClassifier {
+    explore(section) {
+        Array.from(section.children).forEach((primaryChild, i, children) => {
+            // skip last two elements
+            if (i >= section.children.size - 2)
+                return;
+            const secondaryChild = children[i + 1];
+            const tertiaryChild = children[i + 2];
+            const primaryChildHasHouseNumber = primaryChild.is("house_number");
+            const secondaryChildHasStreet = secondaryChild.is("street");
+            const secondaryChildPhraseHasStreet = SectionClassifier.findPhrasesContaining(section, secondaryChild).some((p) => {
+                return p.is("street");
+            });
+            const tertiaryChildHasStreetSuffix = tertiaryChild.is("street_suffix");
+            if (primaryChildHasHouseNumber &&
+                (secondaryChildHasStreet || secondaryChildPhraseHasStreet) &&
+                tertiaryChildHasStreetSuffix) {
+                // Every child must be part of the set above and must not omit any children.
+                const matches = Iterator.from(section.phrases).filter((phrase) => {
+                    if (phrase.children.size !== 3)
+                        return false;
+                    const [primaryPhrase, secondPhrase, tertiaryPhrase] = phrase.children;
+                    return (
+                    // ---
+                    primaryPhrase === primaryChild &&
+                        // ---
+                        secondPhrase === secondaryChild &&
+                        // ---
+                        tertiaryPhrase === tertiaryChild);
+                });
+                for (const match of matches) {
+                    match.classifications.add("adjacent");
+                }
+            }
+        });
+    }
+}
+//# sourceMappingURL=AdjacencyClassifier.js.map

package/out/classifiers/AdjacencyClassifier.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"AdjacencyClassifier.js","sourceRoot":"","sources":["../../classifiers/AdjacencyClassifier.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,iBAAiB,EAAQ,MAAM,gBAAgB,CAAA;AAExD;;;;;GAKG;AACH,MAAM,OAAO,mBAAoB,SAAQ,iBAAiB;IACzD,OAAO,CAAC,OAAa;QACpB,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,OAAO,CAAC,CAAC,YAAY,EAAE,CAAC,EAAE,QAAyB,EAAE,EAAE;YACnF,yBAAyB;YACzB,IAAI,CAAC,IAAI,OAAO,CAAC,QAAQ,CAAC,IAAI,GAAG,CAAC;gBAAE,OAAM;YAE1C,MAAM,cAAc,GAAG,QAAQ,CAAC,CAAC,GAAG,CAAC,CAAE,CAAA;YACvC,MAAM,aAAa,GAAG,QAAQ,CAAC,CAAC,GAAG,CAAC,CAAE,CAAA;YAEtC,MAAM,0BAA0B,GAAG,YAAY,CAAC,EAAE,CAAC,cAAc,CAAC,CAAA;YAClE,MAAM,uBAAuB,GAAG,cAAc,CAAC,EAAE,CAAC,QAAQ,CAAC,CAAA;YAC3D,MAAM,6BAA6B,GAAG,iBAAiB,CAAC,qBAAqB,CAAC,OAAO,EAAE,cAAc,CAAC,CAAC,IAAI,CAC1G,CAAC,CAAC,EAAE,EAAE;gBACL,OAAO,CAAC,CAAC,EAAE,CAAC,QAAQ,CAAC,CAAA;YACtB,CAAC,CACD,CAAA;YAED,MAAM,4BAA4B,GAAG,aAAa,CAAC,EAAE,CAAC,eAAe,CAAC,CAAA;YAEtE,IACC,0BAA0B;gBAC1B,CAAC,uBAAuB,IAAI,6BAA6B,CAAC;gBAC1D,4BAA4B,EAC3B,CAAC;gBACF,4EAA4E;gBAC5E,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE;oBAChE,IAAI,MAAM,CAAC,QAAQ,CAAC,IAAI,KAAK,CAAC;wBAAE,OAAO,KAAK,CAAA;oBAE5C,MAAM,CAAC,aAAa,EAAE,YAAY,EAAE,cAAc,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAA;oBAErE,OAAO;oBACN,MAAM;oBACN,aAAa,KAAK,YAAY;wBAC9B,MAAM;wBACN,YAAY,KAAK,cAAc;wBAC/B,MAAM;wBACN,cAAc,KAAK,aAAa,CAChC,CAAA;gBACF,CAAC,CAAC,CAAA;gBAEF,KAAK,MAAM,KAAK,IAAI,OAAO,EAAE,CAAC;oBAC7B,KAAK,CAAC,eAAe,CAAC,GAAG,CAAC,UAAU,CAAC,CAAA;gBACtC,CAAC;YACF,CAAC;QACF,CAAC,CAAC,CAAA;IACH,CAAC;CACD"}

package/out/classifiers/AlphaNumericClassifier.d.ts

@@ -0,0 +1,10 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { Span, WordClassifier } from "mailwoman/core";
+export declare class AlphaNumericClassifier extends WordClassifier {
+    explore(span: Span): void;
+}
+//# sourceMappingURL=AlphaNumericClassifier.d.ts.map

package/out/classifiers/AlphaNumericClassifier.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"AlphaNumericClassifier.d.ts","sourceRoot":"","sources":["../../classifiers/AlphaNumericClassifier.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,IAAI,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAA;AAErD,qBAAa,sBAAuB,SAAQ,cAAc;IAClD,OAAO,CAAC,IAAI,EAAE,IAAI,GAAG,IAAI;CAWhC"}

package/out/classifiers/AlphaNumericClassifier.js

@@ -0,0 +1,23 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { WordClassifier } from "mailwoman/core";
+export class AlphaNumericClassifier extends WordClassifier {
+    explore(span) {
+        if (span.flags.has("numeric")) {
+            span.classifications.add("numeric");
+        }
+        else if (span.flags.has("numeral")) {
+            span.classifications.add("alphanumeric");
+        }
+        else if (span.flags.has("punctuation")) {
+            span.classifications.add("punctuation");
+        }
+        else {
+            span.classifications.add("alpha");
+        }
+    }
+}
+//# sourceMappingURL=AlphaNumericClassifier.js.map

package/out/classifiers/AlphaNumericClassifier.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"AlphaNumericClassifier.js","sourceRoot":"","sources":["../../classifiers/AlphaNumericClassifier.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAQ,cAAc,EAAE,MAAM,gBAAgB,CAAA;AAErD,MAAM,OAAO,sBAAuB,SAAQ,cAAc;IAClD,OAAO,CAAC,IAAU;QACxB,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC;YAC/B,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,SAAS,CAAC,CAAA;QACpC,CAAC;aAAM,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,EAAE,CAAC;YACtC,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,cAAc,CAAC,CAAA;QACzC,CAAC;aAAM,IAAI,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,aAAa,CAAC,EAAE,CAAC;YAC1C,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,aAAa,CAAC,CAAA;QACxC,CAAC;aAAM,CAAC;YACP,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC,OAAO,CAAC,CAAA;QAClC,CAAC;IACF,CAAC;CACD"}

package/out/classifiers/AlphaNumericClassifier.test.d.ts

@@ -0,0 +1,7 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+export {};
+//# sourceMappingURL=AlphaNumericClassifier.test.d.ts.map

package/out/classifiers/AlphaNumericClassifier.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"AlphaNumericClassifier.test.d.ts","sourceRoot":"","sources":["../../classifiers/AlphaNumericClassifier.test.ts"],"names":[],"mappings":"AAAA;;;;GAIG"}

package/out/classifiers/AlphaNumericClassifier.test.js

@@ -0,0 +1,80 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { ClassificationsMatchMap } from "mailwoman/core";
+import test from "tape";
+import { AlphaNumericClassifier } from "./AlphaNumericClassifier.js";
+const classifier = new AlphaNumericClassifier();
+test("AlphaClassification: English letter", (t) => {
+    const span = classifier.classify("A");
+    t.same(span.classifications, ClassificationsMatchMap.from("alpha"));
+    t.end();
+});
+test("AlphaClassification: English mixed-case word", (t) => {
+    const span = classifier.classify("TesT ExAmPle");
+    t.same(span.classifications, ClassificationsMatchMap.from("alpha"));
+    t.end();
+});
+test("AlphaClassification: Japanese", (t) => {
+    const span = classifier.classify("東京");
+    t.same(span.classifications, ClassificationsMatchMap.from("alpha"));
+    t.end();
+});
+test("AlphaClassification: Mandarin", (t) => {
+    const span = classifier.classify("北京市");
+    t.same(span.classifications, ClassificationsMatchMap.from("alpha"));
+    t.end();
+});
+test("AlphaClassification: Cyrillic", (t) => {
+    const span = classifier.classify("Москва́");
+    t.same(span.classifications, ClassificationsMatchMap.from("alpha"));
+    t.end();
+});
+test("NumericClassification: single digit", (t) => {
+    const span = classifier.classify("1");
+    t.same(span.classifications, ClassificationsMatchMap.from("numeric"));
+    t.end();
+});
+test("NumericClassification: multiple digits", (t) => {
+    const span = classifier.classify("1234567890");
+    t.same(span.classifications, ClassificationsMatchMap.from("numeric"));
+    t.end();
+});
+test("PunctuationClassification: single char", (t) => {
+    const span = classifier.classify("@");
+    t.same(span.classifications, ClassificationsMatchMap.from("punctuation"));
+    t.end();
+});
+test("PunctuationClassification: multiple chars", (t) => {
+    const span = classifier.classify("###&$%");
+    t.same(span.classifications, ClassificationsMatchMap.from("punctuation"));
+    t.end();
+});
+test("AlphaNumericClassification: English letter", (t) => {
+    const span = classifier.classify("1A");
+    t.same(span.classifications, ClassificationsMatchMap.from("alphanumeric"));
+    t.end();
+});
+test("AlphaNumericClassification: English mixed-case word", (t) => {
+    const span = classifier.classify("100 TesT ExAmPle");
+    t.same(span.classifications, ClassificationsMatchMap.from("alphanumeric"));
+    t.end();
+});
+test("AlphaNumericClassification: Japanese", (t) => {
+    const span = classifier.classify("1東京");
+    t.same(span.classifications, ClassificationsMatchMap.from("alphanumeric"));
+    t.end();
+});
+test("AlphaNumericClassification: Mandarin", (t) => {
+    const span = classifier.classify("北京市1");
+    t.same(span.classifications, ClassificationsMatchMap.from("alphanumeric"));
+    t.end();
+});
+test("AlphaNumericClassification: Cyrillic", (t) => {
+    const span = classifier.classify("1Москва́");
+    t.same(span.classifications, ClassificationsMatchMap.from("alphanumeric"));
+    t.end();
+});
+//# sourceMappingURL=AlphaNumericClassifier.test.js.map

package/out/classifiers/AlphaNumericClassifier.test.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"AlphaNumericClassifier.test.js","sourceRoot":"","sources":["../../classifiers/AlphaNumericClassifier.test.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,uBAAuB,EAAE,MAAM,gBAAgB,CAAA;AACxD,OAAO,IAAI,MAAM,MAAM,CAAA;AACvB,OAAO,EAAE,sBAAsB,EAAE,MAAM,6BAA6B,CAAA;AAEpE,MAAM,UAAU,GAAG,IAAI,sBAAsB,EAAE,CAAA;AAE/C,IAAI,CAAC,qCAAqC,EAAE,CAAC,CAAC,EAAE,EAAE;IACjD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAA;IACrC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAA;IACnE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,8CAA8C,EAAE,CAAC,CAAC,EAAE,EAAE;IAC1D,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,cAAc,CAAC,CAAA;IAChD,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAA;IACnE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,+BAA+B,EAAE,CAAC,CAAC,EAAE,EAAE;IAC3C,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAA;IACtC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAA;IACnE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,+BAA+B,EAAE,CAAC,CAAC,EAAE,EAAE;IAC3C,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAA;IACvC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAA;IACnE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,+BAA+B,EAAE,CAAC,CAAC,EAAE,EAAE;IAC3C,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAA;IAC3C,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAA;IACnE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,qCAAqC,EAAE,CAAC,CAAC,EAAE,EAAE;IACjD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAA;IACrC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAA;IACrE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,wCAAwC,EAAE,CAAC,CAAC,EAAE,EAAE;IACpD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAA;IAC9C,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAA;IACrE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,wCAAwC,EAAE,CAAC,CAAC,EAAE,EAAE;IACpD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAA;IACrC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAA;IACzE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,2CAA2C,EAAE,CAAC,CAAC,EAAE,EAAE;IACvD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAA;IAC1C,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAA;IACzE,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,4CAA4C,EAAE,CAAC,CAAC,EAAE,EAAE;IACxD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAA;IACtC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,CAAA;IAC1E,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,qDAAqD,EAAE,CAAC,CAAC,EAAE,EAAE;IACjE,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,kBAAkB,CAAC,CAAA;IACpD,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,CAAA;IAC1E,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,sCAAsC,EAAE,CAAC,CAAC,EAAE,EAAE;IAClD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAA;IACvC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,CAAA;IAC1E,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,sCAAsC,EAAE,CAAC,CAAC,EAAE,EAAE;IAClD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAA;IACxC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,CAAA;IAC1E,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA;AAEF,IAAI,CAAC,sCAAsC,EAAE,CAAC,CAAC,EAAE,EAAE;IAClD,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAA;IAC5C,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,EAAE,uBAAuB,CAAC,IAAI,CAAC,cAAc,CAAC,CAAC,CAAA;IAC1E,CAAC,CAAC,GAAG,EAAE,CAAA;AACR,CAAC,CAAC,CAAA"}

package/out/classifiers/CentralEuropeanStreetNameClassifier.d.ts

@@ -0,0 +1,16 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { SectionClassifier, Span } from "mailwoman/core";
+/**
+ * Classifies street names with no suffix or prefix when accompanied by a house number in the same
+ * section.
+ *
+ * @see {@link https://github.com/pelias/parser/issues/83 | Issue #83}
+ */
+export declare class CentralEuropeanStreetNameClassifier extends SectionClassifier {
+    explore(section: Span): void;
+}
+//# sourceMappingURL=CentralEuropeanStreetNameClassifier.d.ts.map

package/out/classifiers/CentralEuropeanStreetNameClassifier.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"CentralEuropeanStreetNameClassifier.d.ts","sourceRoot":"","sources":["../../classifiers/CentralEuropeanStreetNameClassifier.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,iBAAiB,EAAE,IAAI,EAAE,MAAM,gBAAgB,CAAA;AAExD;;;;;GAKG;AACH,qBAAa,mCAAoC,SAAQ,iBAAiB;IAClE,OAAO,CAAC,OAAO,EAAE,IAAI,GAAG,IAAI;CAsBnC"}

package/out/classifiers/CentralEuropeanStreetNameClassifier.js

@@ -0,0 +1,36 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { SectionClassifier } from "mailwoman/core";
+/**
+ * Classifies street names with no suffix or prefix when accompanied by a house number in the same
+ * section.
+ *
+ * @see {@link https://github.com/pelias/parser/issues/83 | Issue #83}
+ */
+export class CentralEuropeanStreetNameClassifier extends SectionClassifier {
+    explore(section) {
+        const { first, last } = section.children;
+        // Does the section have at least two children?
+        if (!first || !last)
+            return;
+        // Is the first child a toponym, like a region or city?
+        if (first.is("toponym"))
+            return;
+        // Section doesn't end with a housenumber?
+        if (!last.is("house_number"))
+            return;
+        // Other elements cannot contain any public classifications.
+        if (first.classifications.hasVisibleClassification())
+            return;
+        // Assume the first token is a street name...
+        first.classifications.add({
+            classification: "street",
+            confidence: 0.5,
+            flags: new Set(["central_european_street_name"]),
+        });
+    }
+}
+//# sourceMappingURL=CentralEuropeanStreetNameClassifier.js.map

package/out/classifiers/CentralEuropeanStreetNameClassifier.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"CentralEuropeanStreetNameClassifier.js","sourceRoot":"","sources":["../../classifiers/CentralEuropeanStreetNameClassifier.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,iBAAiB,EAAQ,MAAM,gBAAgB,CAAA;AAExD;;;;;GAKG;AACH,MAAM,OAAO,mCAAoC,SAAQ,iBAAiB;IAClE,OAAO,CAAC,OAAa;QAC3B,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,OAAO,CAAC,QAAQ,CAAA;QAExC,+CAA+C;QAC/C,IAAI,CAAC,KAAK,IAAI,CAAC,IAAI;YAAE,OAAM;QAE3B,uDAAuD;QACvD,IAAI,KAAK,CAAC,EAAE,CAAC,SAAS,CAAC;YAAE,OAAM;QAE/B,0CAA0C;QAC1C,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,cAAc,CAAC;YAAE,OAAM;QAEpC,4DAA4D;QAC5D,IAAI,KAAK,CAAC,eAAe,CAAC,wBAAwB,EAAE;YAAE,OAAM;QAE5D,6CAA6C;QAC7C,KAAK,CAAC,eAAe,CAAC,GAAG,CAAC;YACzB,cAAc,EAAE,QAAQ;YACxB,UAAU,EAAE,GAAG;YACf,KAAK,EAAE,IAAI,GAAG,CAAC,CAAC,8BAA8B,CAAC,CAAC;SAChD,CAAC,CAAA;IACH,CAAC;CACD"}

package/out/classifiers/CentralEuropeanStreetNameClassifier.test.d.ts

@@ -0,0 +1,7 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+export {};
+//# sourceMappingURL=CentralEuropeanStreetNameClassifier.test.d.ts.map

package/out/classifiers/CentralEuropeanStreetNameClassifier.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"CentralEuropeanStreetNameClassifier.test.d.ts","sourceRoot":"","sources":["../../classifiers/CentralEuropeanStreetNameClassifier.test.ts"],"names":[],"mappings":"AAAA;;;;GAIG"}

package/out/classifiers/CentralEuropeanStreetNameClassifier.test.js

@@ -0,0 +1,70 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { ClassificationsMatchMap, Span } from "mailwoman/core";
+import test from "tape";
+import { CentralEuropeanStreetNameClassifier } from "./CentralEuropeanStreetNameClassifier.js";
+const classifier = new CentralEuropeanStreetNameClassifier();
+const foo = Span.from("Foo");
+const fooHouseNum = Span.from("1", {
+    start: 4,
+    classifications: ["house_number"],
+});
+foo.nextSiblings.add(fooHouseNum);
+const bar = Span.from("Bar");
+const barHouseNum = Span.from("2137", {
+    start: 4,
+    classifications: ["house_number"],
+});
+bar.nextSiblings.add(barHouseNum);
+const baz = Span.from("Baz");
+const bazHouseNum0 = Span.from("152/160", {
+    start: 4,
+    classifications: ["house_number"],
+});
+const bazHouseNum1 = Span.from("152", {
+    start: 4,
+    classifications: ["house_number"],
+});
+const bazHouseNum2 = Span.from("160", {
+    start: 8,
+    classifications: ["house_number"],
+});
+baz.nextSiblings.add(bazHouseNum0, bazHouseNum1);
+bazHouseNum1.nextSiblings.add(bazHouseNum2);
+// The Qux test case covers when the section has a greater length than
+// the tokens it contains, such as when it ends with whitespace.
+const qux = Span.from("Qux");
+const quxHouseNum = Span.from("1", {
+    start: 4,
+    classifications: ["house_number"],
+});
+qux.nextSiblings.add(quxHouseNum);
+const valid = [
+    Span.from("Foo 1", { children: [foo, fooHouseNum] }),
+    Span.from("Bar 2137", { children: [bar, barHouseNum] }),
+    Span.from("Baz 152/160", { children: [baz, bazHouseNum0, bazHouseNum1, bazHouseNum2] }),
+    Span.from("Qux 1 ", { children: [qux, quxHouseNum] }),
+];
+valid.forEach((span) => {
+    test(`classify: ${span.body}`, (t) => {
+        classifier.explore(span);
+        const [head, ...tail] = span.children;
+        // first child should now be classified as a street
+        t.same(head.classifications, ClassificationsMatchMap.from({
+            classification: "street",
+            confidence: 0.5,
+            flags: new Set(["central_european_street_name"]),
+        }), `'${span.body}'`);
+        tail.forEach((c) => {
+            t.same(c.classifications, ClassificationsMatchMap.from({
+                classification: "house_number",
+                confidence: 1,
+            }), `'${span.body}'`);
+        });
+        t.end();
+    });
+});
+//# sourceMappingURL=CentralEuropeanStreetNameClassifier.test.js.map

package/out/classifiers/CentralEuropeanStreetNameClassifier.test.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"CentralEuropeanStreetNameClassifier.test.js","sourceRoot":"","sources":["../../classifiers/CentralEuropeanStreetNameClassifier.test.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,uBAAuB,EAAE,IAAI,EAAE,MAAM,gBAAgB,CAAA;AAC9D,OAAO,IAAI,MAAM,MAAM,CAAA;AACvB,OAAO,EAAE,mCAAmC,EAAE,MAAM,0CAA0C,CAAA;AAE9F,MAAM,UAAU,GAAG,IAAI,mCAAmC,EAAE,CAAA;AAE5D,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;AAC5B,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE;IAClC,KAAK,EAAE,CAAC;IACR,eAAe,EAAE,CAAC,cAAc,CAAC;CACjC,CAAC,CAAA;AAEF,GAAG,CAAC,YAAY,CAAC,GAAG,CAAC,WAAW,CAAC,CAAA;AAEjC,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;AAC5B,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE;IACrC,KAAK,EAAE,CAAC;IACR,eAAe,EAAE,CAAC,cAAc,CAAC;CACjC,CAAC,CAAA;AACF,GAAG,CAAC,YAAY,CAAC,GAAG,CAAC,WAAW,CAAC,CAAA;AAEjC,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;AAC5B,MAAM,YAAY,GAAG,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE;IACzC,KAAK,EAAE,CAAC;IACR,eAAe,EAAE,CAAC,cAAc,CAAC;CACjC,CAAC,CAAA;AACF,MAAM,YAAY,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,EAAE;IACrC,KAAK,EAAE,CAAC;IACR,eAAe,EAAE,CAAC,cAAc,CAAC;CACjC,CAAC,CAAA;AACF,MAAM,YAAY,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,EAAE;IACrC,KAAK,EAAE,CAAC;IACR,eAAe,EAAE,CAAC,cAAc,CAAC;CACjC,CAAC,CAAA;AAEF,GAAG,CAAC,YAAY,CAAC,GAAG,CAAC,YAAY,EAAE,YAAY,CAAC,CAAA;AAEhD,YAAY,CAAC,YAAY,CAAC,GAAG,CAAC,YAAY,CAAC,CAAA;AAE3C,sEAAsE;AACtE,gEAAgE;AAChE,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;AAC5B,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE;IAClC,KAAK,EAAE,CAAC;IACR,eAAe,EAAE,CAAC,cAAc,CAAC;CACjC,CAAC,CAAA;AACF,GAAG,CAAC,YAAY,CAAC,GAAG,CAAC,WAAW,CAAC,CAAA;AAEjC,MAAM,KAAK,GAAG;IACb,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,EAAE,QAAQ,EAAE,CAAC,GAAG,EAAE,WAAW,CAAC,EAAE,CAAC;IACpD,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,EAAE,QAAQ,EAAE,CAAC,GAAG,EAAE,WAAW,CAAC,EAAE,CAAC;IACvD,IAAI,CAAC,IAAI,CAAC,aAAa,EAAE,EAAE,QAAQ,EAAE,CAAC,GAAG,EAAE,YAAY,EAAE,YAAY,EAAE,YAAY,CAAC,EAAE,CAAC;IACvF,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,EAAE,QAAQ,EAAE,CAAC,GAAG,EAAE,WAAW,CAAC,EAAE,CAAC;CACrD,CAAA;AAED,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;IACtB,IAAI,CAAC,aAAa,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,EAAE;QACpC,UAAU,CAAC,OAAO,CAAC,IAAI,CAAC,CAAA;QAExB,MAAM,CAAC,IAAI,EAAE,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAA;QAErC,mDAAmD;QACnD,CAAC,CAAC,IAAI,CACL,IAAK,CAAC,eAAe,EACrB,uBAAuB,CAAC,IAAI,CAAC;YAC5B,cAAc,EAAE,QAAQ;YACxB,UAAU,EAAE,GAAG;YACf,KAAK,EAAE,IAAI,GAAG,CAAC,CAAC,8BAA8B,CAAC,CAAC;SAChD,CAAC,EACF,IAAI,IAAI,CAAC,IAAI,GAAG,CAChB,CAAA;QAED,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE;YAClB,CAAC,CAAC,IAAI,CACL,CAAC,CAAC,eAAe,EACjB,uBAAuB,CAAC,IAAI,CAAC;gBAC5B,cAAc,EAAE,cAAc;gBAC9B,UAAU,EAAE,CAAC;aACb,CAAC,EACF,IAAI,IAAI,CAAC,IAAI,GAAG,CAChB,CAAA;QACF,CAAC,CAAC,CAAA;QAEF,CAAC,CAAC,GAAG,EAAE,CAAA;IACR,CAAC,CAAC,CAAA;AACH,CAAC,CAAC,CAAA"}

package/out/classifiers/ChainClassifier.d.ts

@@ -0,0 +1,14 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { PhraseClassifier, Span } from "mailwoman/core";
+/**
+ * Chain-restaurants and other franchises.
+ */
+export declare class ChainClassifier extends PhraseClassifier {
+    ready(): Promise<this>;
+    explore(span: Span): void;
+}
+//# sourceMappingURL=ChainClassifier.d.ts.map

package/out/classifiers/ChainClassifier.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ChainClassifier.d.ts","sourceRoot":"","sources":["../../classifiers/ChainClassifier.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,gBAAgB,EAAE,IAAI,EAAsB,MAAM,gBAAgB,CAAA;AAE3E;;GAEG;AACH,qBAAa,eAAgB,SAAQ,gBAAgB;IACvC,KAAK,IAAI,OAAO,CAAC,IAAI,CAAC;IAM5B,OAAO,CAAC,IAAI,EAAE,IAAI,GAAG,IAAI;CAUhC"}

package/out/classifiers/ChainClassifier.js

@@ -0,0 +1,25 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { PhraseClassifier, prepareLocaleIndex } from "mailwoman/core";
+/**
+ * Chain-restaurants and other franchises.
+ */
+export class ChainClassifier extends PhraseClassifier {
+    async ready() {
+        this.index = await prepareLocaleIndex(this.languages, "chains.txt");
+        return this;
+    }
+    explore(span) {
+        const languages = this.index.get(span.normalized);
+        if (languages) {
+            span.classifications.add({
+                classification: "chain",
+                languages,
+            });
+        }
+    }
+}
+//# sourceMappingURL=ChainClassifier.js.map

package/out/classifiers/ChainClassifier.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ChainClassifier.js","sourceRoot":"","sources":["../../classifiers/ChainClassifier.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,gBAAgB,EAAQ,kBAAkB,EAAE,MAAM,gBAAgB,CAAA;AAE3E;;GAEG;AACH,MAAM,OAAO,eAAgB,SAAQ,gBAAgB;IAC7C,KAAK,CAAC,KAAK;QACjB,IAAI,CAAC,KAAK,GAAG,MAAM,kBAAkB,CAAC,IAAI,CAAC,SAAS,EAAE,YAAY,CAAC,CAAA;QAEnE,OAAO,IAAI,CAAA;IACZ,CAAC;IAEM,OAAO,CAAC,IAAU;QACxB,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,CAAA;QAEjD,IAAI,SAAS,EAAE,CAAC;YACf,IAAI,CAAC,eAAe,CAAC,GAAG,CAAC;gBACxB,cAAc,EAAE,OAAO;gBACvB,SAAS;aACT,CAAC,CAAA;QACH,CAAC;IACF,CAAC;CACD"}

package/out/classifiers/ChainClassifier.test.d.ts

@@ -0,0 +1,7 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+export {};
+//# sourceMappingURL=ChainClassifier.test.d.ts.map

package/out/classifiers/ChainClassifier.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ChainClassifier.test.d.ts","sourceRoot":"","sources":["../../classifiers/ChainClassifier.test.ts"],"names":[],"mappings":"AAAA;;;;GAIG"}

package/out/classifiers/ChainClassifier.test.js

@@ -0,0 +1,15 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+import { assertClassification } from "mailwoman/sdk/test";
+import { ChainClassifier } from "./ChainClassifier.js";
+const classifier = await new ChainClassifier().ready();
+assertClassification(classifier, "chain", [
+    ["McDonalds", ["all"]],
+    ["McDonald's", ["all"]],
+    ["lone star steakhouse", ["all"]],
+    ["panda express", ["all"]],
+]);
+//# sourceMappingURL=ChainClassifier.test.js.map

package/out/classifiers/ChainClassifier.test.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ChainClassifier.test.js","sourceRoot":"","sources":["../../classifiers/ChainClassifier.test.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,OAAO,EAAE,oBAAoB,EAAE,MAAM,oBAAoB,CAAA;AACzD,OAAO,EAAE,eAAe,EAAE,MAAM,sBAAsB,CAAA;AAEtD,MAAM,UAAU,GAAG,MAAM,IAAI,eAAe,EAAE,CAAC,KAAK,EAAE,CAAA;AAEtD,oBAAoB,CAAC,UAAU,EAAE,OAAO,EAAE;IACzC,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,CAAC;IACtB,CAAC,YAAY,EAAE,CAAC,KAAK,CAAC,CAAC;IACvB,CAAC,sBAAsB,EAAE,CAAC,KAAK,CAAC,CAAC;IACjC,CAAC,eAAe,EAAE,CAAC,KAAK,CAAC,CAAC;CAC1B,CAAC,CAAA"}

package/out/classifiers/CompositeClassifier.test.d.ts

@@ -0,0 +1,7 @@
+/**
+ * @copyright Sister Software
+ * @license AGPL-3.0
+ * @author Teffen Ellis, et al.
+ */
+export {};
+//# sourceMappingURL=CompositeClassifier.test.d.ts.map

package/out/classifiers/CompositeClassifier.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"CompositeClassifier.test.d.ts","sourceRoot":"","sources":["../../classifiers/CompositeClassifier.test.ts"],"names":[],"mappings":"AAAA;;;;GAIG"}