First commit

app.py

@@ -0,0 +1,107 @@
+import pathlib
+import os
+import xml.etree.ElementTree as etree
+
+def check_sentence_id(sentence_id):
+    assert(sentence_id.startswith("GF"))
+    assert(len(sentence_id.split(".")) == 3)
+
+def get_file_contents(sentence_id):
+    word_base_id = sentence_id.split(".")[0]
+    gf_folder = os.environ["GF_FOLDER"]
+
+    gf_folder = pathlib.Path(gf_folder)
+    gf_inner_folder = gf_folder.joinpath(word_base_id[:4])
+    gf_file = gf_inner_folder.joinpath(word_base_id + "-dedup.xml")
+
+    assert(gf_file.exists())
+
+    with gf_file.open('r') as fp:
+        return fp.read()
+
+def obtain_sentence(sentence_id, content):
+    substring = '<s xml:id=\"' + sentence_id
+
+    idx_start = content.index(substring)
+    idx_end = content.index("</s>", idx_start) + 4
+
+    sentence = content[idx_start:idx_end]
+    sentence = sentence.replace(" xml:", " ")
+
+    return sentence
+
+def as_example(sentence_id, headword, sentence):
+    root = etree.XML(sentence)
+    result = {}
+    current_text = ""
+
+    for element in root:
+        if element.text == headword:
+            # first, lets pack left part
+            result["left"] = current_text
+            # then, lets pack the mid - headword part
+            result["mid"] = element.text
+            # and cleanup the current_text
+            current_text = ""
+
+        else:
+            # else, lets just add
+            current_text += element.text
+
+    # lastly add the right part
+    if "mid" not in result:
+        return "headword not found"
+    elif "left" not in result:
+        result["left"] = ""
+
+    result["right"] = current_text
+    return result
+
+def obtain_example_from_gf2(sentence_id, headword):
+    try:
+        check_sentence_id(sentence_id)
+        contents = get_file_contents(sentence_id)
+        sentence = obtain_sentence(sentence_id, contents)
+        return as_example(sentence_id, headword, sentence)
+    except ValueError:
+        return "Could not find word id"
+    except IndexError:
+        return "Word id index after last dot should be an integer"
+    except AssertionError:
+        return "Bad word id"
+    except KeyError:
+        return "GF2 location not set on the server"
+
+
+from flask import Response, Flask, request, jsonify
+app = Flask(__name__)
+
+@app.route('/')
+def home():
+    return "An api to get examples from gigafida, just use /get_example/$WORD_ID to get them"
+
+@app.route('/get_example/<string:sentence_id>/<string:headword>')
+def test(sentence_id, headword):
+    result = obtain_example_from_gf2(sentence_id, headword)
+    mimetype = "application/json" if type(result) is dict else "text/plain"
+    return Response(result, mimetype=mimetype)
+
+@app.route('/get_examples', methods=["POST"])
+def get_examples():
+    data = request.get_json()
+    if data is None or type(data) is not dict:
+        return "", 500
+
+    bad_ones = {}
+    good_ones = {}
+    for sentence_id, headword in data.items():
+        headword = headword.strip()
+        example = obtain_example_from_gf2(sentence_id, headword)
+        if type(example) is dict:
+            good_ones[sentence_id] = example
+        else:
+            bad_ones[sentence_id] = example
+
+    return jsonify({"good": good_ones, "bad": bad_ones})
+
+

run.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+# cd to script dir
+cd "$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
+
+# activate python venv
+source ./venv/bin/activate
+
+# set gf2 location
+export GF_FOLDER="/net/nas/resources/corpus/gigafida/v2.0/gigafida_dedup/tei/data/"
+
+# run the app
+flask run --port=2001 --host=0.0.0.0 --debugger --reload