2f:T27a2,default_pairwise_comparison_data = [ { match_weight: 24.72661262857488, match_probability: 0.9999999639796531, source_dataset_l: "_a", source_dataset_r: "_a", unique_id_l: 79, unique_id_r: 79, first_name_l: "Carr", first_name_r: "Carr", gamma_first_name: 3, tf_first_name_l: 0.0033333333333333335, tf_first_name_r: 0.0033333333333333335, bf_first_name: 185.00697674418603, bf_tf_adj_first_name: 1.540482445665154, surname_l: "Ethan ", surname_r: "Ethan ", gamma_surname: 3, tf_surname_l: 0.005417118093174431, tf_surname_r: 0.005417118093174431, bf_surname: 130.37728776185224, bf_tf_adj_surname: 1.3450962434525533, dob_l: "2013-01-21", dob_r: "2013-01-21", gamma_dob: 3, tf_dob_l: 0.006, tf_dob_r: 0.006, bf_dob: 284.4990909090909, bf_tf_adj_dob: 0.5565337057049764, email_l: "stacyball@medina.biz", email_r: "stacyball@medina.biz", gamma_email: 2, tf_email_l: 0.009009009009009009, tf_email_r: 0.009009009009009009, bf_email: 257.99222737819025, bf_tf_adj_email: 0.4087332439105659, city_l: "London", city_r: "London", gamma_city: 1, tf_city_l: 0.2884399551066218, tf_city_r: 0.2884399551066218, bf_city: 10.263374411018484, bf_tf_adj_city: 0.32090613034647414 }, { match_weight: 11.543499994820714, match_probability: 0.9996651002856178, source_dataset_l: "_b", source_dataset_r: "_a", unique_id_l: 80, unique_id_r: 79, first_name_l: "Carr", first_name_r: "Carr", gamma_first_name: 3, tf_first_name_l: 0.0033333333333333335, tf_first_name_r: 0.0033333333333333335, bf_first_name: 185.00697674418603, bf_tf_adj_first_name: 1.540482445665154, surname_l: "Ethan ", surname_r: "Ethan ", gamma_surname: 3, tf_surname_l: 0.005417118093174431, tf_surname_r: 0.005417118093174431, bf_surname: 130.37728776185224, bf_tf_adj_surname: 1.3450962434525533, dob_l: "2012-11-15", dob_r: "2013-01-21", gamma_dob: 0, tf_dob_l: 0.001, tf_dob_r: 0.006, bf_dob: 0.01702393748966431, bf_tf_adj_dob: 1.0, email_l: "stacyball@medina.biz", email_r: "stacyball@medina.biz", gamma_email: 2, tf_email_l: 0.009009009009009009, tf_email_r: 0.009009009009009009, bf_email: 257.99222737819025, bf_tf_adj_email: 0.4087332439105659, city_l: "London", city_r: "London", gamma_city: 1, tf_city_l: 0.2884399551066218, tf_city_r: 0.2884399551066218, bf_city: 10.263374411018484, bf_tf_adj_city: 0.32090613034647414 }, { match_weight: -2.599230043972054, match_probability: 0.14165045458783695, source_dataset_l: "_b", source_dataset_r: "_a", unique_id_l: 82, unique_id_r: 79, first_name_l: "Eh na", first_name_r: "Carr", gamma_first_name: 0, tf_first_name_l: 0.0011111111111111111, tf_first_name_r: 0.0033333333333333335, bf_first_name: 0.017349943911181373, bf_tf_adj_first_name: 1.0, surname_l: "Carr", surname_r: "Ethan ", gamma_surname: 0, tf_surname_l: 0.0065005417118093175, tf_surname_r: 0.005417118093174431, bf_surname: 0.01712394347596409, bf_tf_adj_surname: 1.0, dob_l: "2013-01-21", dob_r: "2013-01-21", gamma_dob: 3, tf_dob_l: 0.006, tf_dob_r: 0.006, bf_dob: 284.4990909090909, bf_tf_adj_dob: 0.5565337057049764, email_l: "stacyball@medina.biz", email_r: "stacyball@medina.biz", gamma_email: 2, tf_email_l: 0.009009009009009009, tf_email_r: 0.009009009009009009, bf_email: 257.99222737819025, bf_tf_adj_email: 0.4087332439105659, city_l: "London", city_r: "London", gamma_city: 1, tf_city_l: 0.2884399551066218, tf_city_r: 0.2884399551066218, bf_city: 10.263374411018484, bf_tf_adj_city: 0.32090613034647414 }, { match_weight: 11.543499994820714, match_probability: 0.9996651002856178, source_dataset_l: "_a", source_dataset_r: "_a", unique_id_l: 84, unique_id_r: 79, first_name_l: "Carr", first_name_r: "Carr", gamma_first_name: 3, tf_first_name_l: 0.0033333333333333335, tf_first_name_r: 0.0033333333333333335, bf_first_name: 185.00697674418603, bf_tf_adj_first_name: 1.540482445665154, surname_l: "Ethan ", surname_r: "Ethan ", gamma_surname: 3, tf_surname_l: 0.005417118093174431, tf_surname_r: 0.005417118093174431, bf_surname: 130.37728776185224, bf_tf_adj_surname: 1.3450962434525533, dob_l: "2012-10-21", dob_r: "2013-01-21", gamma_dob: 0, tf_dob_l: 0.001, tf_dob_r: 0.006, bf_dob: 0.01702393748966431, bf_tf_adj_dob: 1.0, email_l: "stacyball@medina.biz", email_r: "stacyball@medina.biz", gamma_email: 2, tf_email_l: 0.009009009009009009, tf_email_r: 0.009009009009009009, bf_email: 257.99222737819025, bf_tf_adj_email: 0.4087332439105659, city_l: "London", city_r: "London", gamma_city: 1, tf_city_l: 0.2884399551066218, tf_city_r: 0.2884399551066218, bf_city: 10.263374411018484, bf_tf_adj_city: 0.32090613034647414 }, { match_weight: -2.599230043972054, match_probability: 0.14165045458783695, source_dataset_l: "_b", source_dataset_r: "_a", unique_id_l: 81, unique_id_r: 79, first_name_l: "Etnah", first_name_r: "Carr", gamma_first_name: 0, tf_first_name_l: 0.0011111111111111111, tf_first_name_r: 0.0033333333333333335, bf_first_name: 0.017349943911181373, bf_tf_adj_first_name: 1.0, surname_l: "Carr", surname_r: "Ethan ", gamma_surname: 0, tf_surname_l: 0.0065005417118093175, tf_surname_r: 0.005417118093174431, bf_surname: 0.01712394347596409, bf_tf_adj_surname: 1.0, dob_l: "2013-01-21", dob_r: "2013-01-21", gamma_dob: 3, tf_dob_l: 0.006, tf_dob_r: 0.006, bf_dob: 284.4990909090909, bf_tf_adj_dob: 0.5565337057049764, email_l: "stacyball@medina.biz", email_r: "stacyball@medina.biz", gamma_email: 2, tf_email_l: 0.009009009009009009, tf_email_r: 0.009009009009009009, bf_email: 257.99222737819025, bf_tf_adj_email: 0.4087332439105659, city_l: "London", city_r: "London", gamma_city: 1, tf_city_l: 0.2884399551066218, tf_city_r: 0.2884399551066218, bf_city: 10.263374411018484, bf_tf_adj_city: 0.32090613034647414 }, { match_weight: -8.500685372099381, match_probability: 0.0027532228214566905, source_dataset_l: "_b", source_dataset_r: "_a", unique_id_l: 86, unique_id_r: 79, first_name_l: "Ethan ", first_name_r: "Carr", gamma_first_name: 0, tf_first_name_l: 0.008888888888888889, tf_first_name_r: 0.0033333333333333335, bf_first_name: 0.017349943911181373, bf_tf_adj_first_name: 1.0, surname_l: "Crr", surname_r: "Ethan ", gamma_surname: 0, tf_surname_l: 0.0010834236186348862, tf_surname_r: 0.005417118093174431, bf_surname: 0.01712394347596409, bf_tf_adj_surname: 1.0, dob_l: "2013-01-21", dob_r: "2013-01-21", gamma_dob: 3, tf_dob_l: 0.006, tf_dob_r: 0.006, bf_dob: 284.4990909090909, bf_tf_adj_dob: 0.5565337057049764, email_l: "stacyball@medina.biz", email_r: "stacyball@medina.biz", gamma_email: 2, tf_email_l: 0.009009009009009009, tf_email_r: 0.009009009009009009, bf_email: 257.99222737819025, bf_tf_adj_email: 0.4087332439105659, city_l: "Londo", city_r: "London", gamma_city: 0, tf_city_l: 0.004489337822671156, tf_city_r: 0.2884399551066218, bf_city: 0.05510019225081342, bf_tf_adj_city: 1.0 }, { match_weight: -2.6328441608207083, match_probability: 0.1388411501653245, source_dataset_l: "_b", source_dataset_r: "_a", unique_id_l: 85, unique_id_r: 79, first_name_l: "Ethan ", first_name_r: "Carr", gamma_first_name: 0, tf_first_name_l: 0.008888888888888889, tf_first_name_r: 0.0033333333333333335, bf_first_name: 0.017349943911181373, bf_tf_adj_first_name: 1.0, surname_l: "", surname_r: "Ethan ", gamma_surname: -1, tf_surname_l: "", tf_surname_r: 0.005417118093174431, bf_surname: 1.0, bf_tf_adj_surname: 1.0, dob_l: "2013-01-21", dob_r: "2013-01-21", gamma_dob: 3, tf_dob_l: 0.006, tf_dob_r: 0.006, bf_dob: 284.4990909090909, bf_tf_adj_dob: 0.5565337057049764, email_l: "stacyball@medina.biz", email_r: "stacyball@medina.biz", gamma_email: 2, tf_email_l: 0.009009009009009009, tf_email_r: 0.009009009009009009, bf_email: 257.99222737819025, bf_tf_adj_email: 0.4087332439105659, city_l: "oodLn", city_r: "London", gamma_city: 0, tf_city_l: 0.002244668911335578, tf_city_r: 0.2884399551066218, bf_city: 0.05510019225081342, bf_tf_adj_city: 1.0 }, { match_weight: -9.319645328878384, match_probability: 0.0015625289718773487, source_dataset_l: "_b", source_dataset_r: "_a", unique_id_l: 83, unique_id_r: 79, first_name_l: "Ethan ", first_name_r: "Carr", gamma_first_name: 0, tf_first_name_l: 0.008888888888888889, tf_first_name_r: 0.0033333333333333335, bf_first_name: 0.017349943911181373, bf_tf_adj_first_name: 1.0, surname_l: "Cra", surname_r: "Ethan ", gamma_surname: 0, tf_surname_l: 0.0010834236186348862, tf_surname_r: 0.005417118093174431, bf_surname: 0.01712394347596409, bf_tf_adj_surname: 1.0, dob_l: "2013-01-21", dob_r: "2013-01-21", gamma_dob: 3, tf_dob_l: 0.006, tf_dob_r: 0.006, bf_dob: 284.4990909090909, bf_tf_adj_dob: 0.5565337057049764, email_l: "", email_r: "stacyball@medina.biz", gamma_email: -1, tf_email_l: "", tf_email_r: 0.009009009009009009, bf_email: 1.0, bf_tf_adj_email: 1.0, city_l: "London", city_r: "London", gamma_city: 1, tf_city_l: 0.2884399551066218, tf_city_r: 0.2884399551066218, bf_city: 10.263374411018484, bf_tf_adj_city: 0.32090613034647414 } ]30:T1732,default_splink_settings = { return { probability_two_random_records_match: 0.01, link_type: "link_only", blocking_rules_to_generate_predictions: [ "l.first_name = r.first_name", "l.surname = r.surname" ], comparisons: [ { output_column_name: "first_name", comparison_levels: [ { sql_condition: '"first_name_l" IS NULL OR "first_name_r" IS NULL', label_for_charts: "Null", is_null_level: true }, { sql_condition: '"first_name_l" = "first_name_r"', label_for_charts: "Exact match", u_probability: 0.0051349414855505135, tf_adjustment_column: "first_name", tf_adjustment_weight: 1.0 }, { sql_condition: 'jaro_winkler_similarity("first_name_l", "first_name_r") >= 0.9', label_for_charts: "Jaro_winkler_similarity >= 0.9", u_probability: 0.002763656214814562 }, { sql_condition: 'jaro_winkler_similarity("first_name_l", "first_name_r") >= 0.7', label_for_charts: "Jaro_winkler_similarity >= 0.7", u_probability: 0.03148350336074243 }, { sql_condition: "ELSE", label_for_charts: "All other comparisons", u_probability: 0.9606178989388925 } ], comparison_description: "Exact match vs. First_Name within jaro_winkler_similarity thresholds 0.9, 0.7 vs. anything else" }, { output_column_name: "surname", comparison_levels: [ { sql_condition: '"surname_l" IS NULL OR "surname_r" IS NULL', label_for_charts: "Null", is_null_level: true }, { sql_condition: '"surname_l" = "surname_r"', label_for_charts: "Exact match", u_probability: 0.007286545197467785, tf_adjustment_column: "surname", tf_adjustment_weight: 1.0 }, { sql_condition: 'jaro_winkler_similarity("surname_l", "surname_r") >= 0.9', label_for_charts: "Jaro_winkler_similarity >= 0.9", u_probability: 0.002643079790481699 }, { sql_condition: 'jaro_winkler_similarity("surname_l", "surname_r") >= 0.7', label_for_charts: "Jaro_winkler_similarity >= 0.7", u_probability: 0.016774317940807866 }, { sql_condition: "ELSE", label_for_charts: "All other comparisons", u_probability: 0.9732960570712427 } ], comparison_description: "Exact match vs. Surname within jaro_winkler_similarity thresholds 0.9, 0.7 vs. anything else" }, { output_column_name: "dob", comparison_levels: [ { sql_condition: '"dob_l" IS NULL OR "dob_r" IS NULL', label_for_charts: "Null", is_null_level: true }, { sql_condition: '"dob_l" = "dob_r"', label_for_charts: "Exact match", u_probability: 0.0033392022342298585, tf_adjustment_column: "dob", tf_adjustment_weight: 1.0 }, { sql_condition: 'levenshtein("dob_l", "dob_r") <= 1', label_for_charts: "Levenshtein <= 1", u_probability: 0.001403139524686486 }, { sql_condition: 'levenshtein("dob_l", "dob_r") <= 2', label_for_charts: "Levenshtein <= 2", u_probability: 0.016244038343485857 }, { sql_condition: "ELSE", label_for_charts: "All other comparisons", u_probability: 0.9790136198975978 } ], comparison_description: "Exact match vs. Dob within levenshtein thresholds 1, 2 vs. anything else" }, { output_column_name: "email", comparison_levels: [ { sql_condition: '"email_l" IS NULL OR "email_r" IS NULL', label_for_charts: "Null", is_null_level: true }, { sql_condition: '"email_l" = "email_r"', label_for_charts: "Exact match", u_probability: 0.0036822814766717645, tf_adjustment_column: "email", tf_adjustment_weight: 1.0 }, { sql_condition: 'levenshtein("email_l", "email_r") <= 2', label_for_charts: "Levenshtein <= 2", u_probability: 0.00035028663699197756 }, { sql_condition: "ELSE", label_for_charts: "All other comparisons", u_probability: 0.9959674318863363 } ], comparison_description: "Exact match vs. Email within levenshtein threshold 2 vs. anything else" }, { output_column_name: "city", comparison_levels: [ { sql_condition: '"city_l" IS NULL OR "city_r" IS NULL', label_for_charts: "Null", is_null_level: true }, { sql_condition: '"city_l" = "city_r"', label_for_charts: "Exact match", u_probability: 0.09256214983057671, tf_adjustment_column: "city", tf_adjustment_weight: 1.0 }, { sql_condition: "ELSE", label_for_charts: "All other comparisons", u_probability: 0.9074378501694232 } ], comparison_description: "Exact match vs. anything else" } ], retain_matching_columns: true, retain_intermediate_calculation_columns: true, max_iterations: 10, em_convergence: 0.01, sql_dialect: "duckdb", linker_uid: "qBuHfYYN" }; }31:T550,draw_header = function (table_element, data_with_inputs_clone, cols) { // Draw the two table header rows let orig_row_clone = slt.cloneDeep(original_row); let first_keys = { clerical_match_score: null, notes_input: null }; orig_row_clone = { ...first_keys, ...orig_row_clone }; const table_first_header = table_element.append("tr"); d3.select(table_first_header.node()) .selectAll("th") .data(colspan_spec) .enter() .append("th") .attr("colspan", (col) => col.span) .text((col) => col.text) .attr("class", (d) => `${d.css_class} table_header table_first_header`); const table_second_header = table_element.append("tr"); const column_header_translations = { clerical_match_score: "Clerical Probability", notes_input: "Notes", match_probability: "Match Probability", match_weight: "Match Weight", model_truth_status: "Label category" }; d3.select(table_second_header.node()) .selectAll("th") .data(cols) .enter() .filter((d) => splink_scores_filter({ _key: d })) .append("th") .attr( "class", (d, i) => `${styling_lookup[i].css_class} table_header table_second_header ` ) .text((col) => { if (col in column_header_translations) { return column_header_translations[col]; } else { return col; } }); }32:T601,render_tooltip_contents = function (tooltip, cell_data, i) { tooltip .html(() => { let col_name = cell_data["_key"]; // Original row is the first (yellow highlighted) row in the table let original_value = original_row[col_name]; let comparison_value = cell_data["_orig_data"][col_name]; let lev = leven(original_value, comparison_value, false); let jw = jaro_winkler_distance(original_value, comparison_value); jw = match_prob_number_format(jw); let diff_view; if (lev > 0) { diff_view = insert_del_differences(comparison_value, original_value); } else { diff_view = ""; } return `

Original:	${original_value}
Comparison:	${comparison_value}
Diff:	${diff_view}
Levenshtein:	${lev}
Jaro winkler:	${jw}

`; }) .style("visibility", "visible"); }33:T4ff,draw_original_row = function (table) { // Draw the top yellow highlighted row let first_cols = { clerical_match_score: null, notes_input: null }; let orig_row_clone = slt.cloneDeep(original_row); orig_row_clone = { ...first_cols, ...orig_row_clone }; orig_row_clone["match_probability"] = match_prob_number_format( orig_row_clone["match_probability"] ); orig_row_clone["match_weight"] = match_prob_number_format( orig_row_clone["match_weight"] ); orig_row_clone["clerical_match_score"] = "Original record"; orig_row_clone["notes_input"] = "(Scores in this row are for an exact match on all columns)"; let orig_row_data = [orig_row_clone]; const table_original_row = table .selectAll("tr.original_record") .data(orig_row_data) .join("tr") .attr("class", "original_record_row"); table_original_row .selectAll("td") .data((d) => Object.entries(d).filter((entry) => splink_scores_filter({ _key: entry[0] }) ) ) .join("td") .text((d) => d[1]) .attr("class", (d, i) => { let this_class = `${styling_lookup[i].css_class} original_record_cell col_${d[0]}`; if (i == 2) { this_class += " orig_record_text_with_arrow_cell"; } return this_class; }); }34:T7ad,draw_potential_matches_rows = function (table, data_with_inputs_clone, el) { const table_rows = table .selectAll("tr.potential_match_table_row") .data(data_with_inputs_clone) .join("tr") .attr("class", "potential_match_table_row"); const table_cells = table_rows .selectAll("td") .data(table_cells_d3_data_mapper) .join("td") .style("overflow", "hidden") .attr("class", (d, i) => { let this_class = `col_${d._key}`; let is_comparison_data_col = columns_to_highlight_similarity.includes( d["_key"] ); if (is_comparison_data_col) { this_class += " potential_match_comparison_data_cell"; } this_class += ` ${styling_lookup[i].css_class}`; return this_class; }); // Draw the table cells by appending into the td cells we just created table_cells.append((original_cell_data) => { let td_contents = original_cell_data["_td_contents"]; let key = original_cell_data["_key"]; // If it's a html element, then bind an oninput event // that updates the value // Then calls a redraw function that is responsible for the visual appearance if (td_contents instanceof HTMLElement) { td_contents.oninput = (event) => { form_oninput_update_value(original_cell_data, el); redraw_potential_match_cells(data_with_inputs_clone, table, el); }; return td_contents; } else { return slt.d3.create("p").text(td_contents).node(); } }); table_cells .filter((d, i) => { return columns_to_highlight_similarity.includes(d["_key"]); }) .on("mouseover", function (event, d, i) { render_tooltip_contents(tooltip, d, i); }) .on("mousemove", function (event) { tooltip .style("top", event.pageY - 10 + "px") .style("left", event.pageX + 10 + "px"); }) .on("mouseout", function (event) { tooltip.html(``).style("visibility", "hidden"); }); }35:Tf1d,styles = { // return html` // `; return html` `; }36:T419,viewof existing_labels_csv = { const el = html``; const button = html` `; el.appendChild(button); el.value = null; function readFileAsText(file) { return new Promise((resolve, reject) => { const reader = new FileReader(); reader.onload = () => resolve(reader.result); reader.onerror = () => reject(reader.error); reader.readAsText(file); }); } // Add an event listener for the "change" event on the button button.addEventListener("change", async (event) => { // Get the selected file const file = event.target.files.item(0); try { // Read the file contents as ArrayBuffer const fileData = await readFileAsText(file); el.value = slt.d3.csvParse(fileData); el.dispatchEvent(new Event("input", { bubbles: true })); } catch (error) { console.error("Error reading file:", error); } }); button.classList.add("labelling-tool-checkbox"); return el; }37:T574,viewof table_interface = { let data_with_inputs_clone = slt.cloneDeep(data_with_inputs); const wrapper = slt.d3.create("div").style("overflow-y", "auto"); // interface is the html node that will be returned and displayed by the interface const interface_node = wrapper.node(); // interface.value is the (reactive viewof) value that's exposed by the interface // i.e. it contains the output data of the interface interface_node.value = {}; // Set the value of el.value to a dict like // {'uid_1': row_data, 'uid_2': row_data} data_with_inputs_clone.forEach((d) => { let uid = d[unique_id_column_name]; d = slt.cloneDeep(d); d["clerical_match_score_value"] = d["clerical_match_score"].value; d["notes_input_value"] = d["notes_input"].value; interface_node.value[uid] = d; }); const table = wrapper.append("table").attr("id", "main-interface-table"); let cols = Object.keys(data_with_inputs_clone[0]); cols = cols.filter((d) => d != "_row_num"); draw_header(table, data_with_inputs_clone, cols); draw_original_row(table); draw_potential_matches_spacer_row(table); draw_potential_matches_rows(table, data_with_inputs_clone, interface_node); // Trigger an update to reflect the current data data_with_inputs_clone[0]["clerical_match_score"].dispatchEvent( new Event("input", { bubbles: true }) ); return interface_node; }38:Tf1f,shade_potential_match_cells = function (d3_data, sel, el) { // The cell additional cell styling // that gives a visual representation of the similarity let col_name = d3_data["_key"]; let cell_value = d3_data["_td_contents"]; let original_value = original_row[col_name]; let row_id = d3_data["_row_uid"]; // What values constitute transitive matches? let potential_match_rows = Object.values(el.value); let transitive_values = []; potential_match_rows.forEach((p_row) => { if (p_row["unique_id"] != row_id) { if (p_row["clerical_match_score_value"] > 0.5) { if (p_row[col_name] != "") { transitive_values.push(p_row[col_name]); } } } }); if (d3_data["_key"] == "clerical_match_score") { sel.style("background-color", (d) => { return white_green_colour_scale_match_prob(d["_td_contents"].value); }); } if (columns_to_highlight_similarity.includes(d3_data["_key"])) { // Direct exact matches if (cell_value == original_value) { sel.style("background-color", (d) => { return white_green_colour_scale(1.0); }); } else { // Compute leven similarity let direct_leven_perc = leven(cell_value, original_value); sel.style("background-color", (d) => { return white_green_colour_scale(direct_leven_perc); }); } // If not exact match look for transitive matches if (cell_value != original_value) { if (transitive_values.includes(cell_value)) { sel.style("border-color", (d) => { return white_blue_colour_scale(1.0); }); sel.style("border-width", transitive_highlight_width); sel.style("padding", padding_with_highlight); sel.style("border-style", "solid"); } else { // Indirect leven matches // Only show if the transitive leven exceeds the direct level let direct_leven_perc = leven(cell_value, original_value); let transitive_leven_perc = 0; if (transitive_leven_perc > direct_leven_perc) { transitive_values.forEach((transitive_value) => { let lev = leven(cell_value, transitive_value); transitive_leven_perc = Math.max(transitive_leven_perc, lev); }); } if (transitive_leven_perc > 0.5) { sel.style("border-width", transitive_highlight_width); sel.style("padding", padding_with_highlight); sel.style("border-style", "solid"); sel.style("border-color", (d) => { return white_blue_colour_scale(transitive_leven_perc); }); } else { sel.style("border-width", null); sel.style("padding", null); sel.style("border-style", null); sel.style("border-color", null); } } } } if (["match_probability", "match_weight"].includes(d3_data["_key"])) { sel.style("background-color", (d) => { let mp = d["_orig_data"]["match_probability"]; mp = parseFloat(mp); return white_green_colour_scale_match_prob(mp); }); } if (d3_data["_key"] == "model_truth_status") { sel.select("p").html((d) => { let data = d["_orig_data"]; let clerical_score = data["clerical_match_score"].value; let splink_score = parseFloat(data["match_probability"]); return score_to_truth_status(clerical_score, splink_score); }); } if (d3_data["_key"] == "model_truth_status") { sel.style("background-color", (d) => { let data = d["_orig_data"]; let clerical_score = data["clerical_match_score"].value; let splink_score = parseFloat(data["match_probability"]); let truth_status = score_to_truth_status(clerical_score, splink_score); if (["FN", "FP"].includes(truth_status)) { return "#ff6d6d"; } else { return null; } }); } }39:T636,redraw_potential_match_cells = function (data_with_inputs_clone, table, el) { // Controls the visual appearance of the potential matches // Does two things: // (1). Highlights which values match the original record // (2). Highlight transitive matches, dependent on the value of the range inputs let sel = table .selectAll("tr.potential_match_table_row") .data(data_with_inputs_clone); const table_rows = sel.join("tr"); const table_cells = table_rows .selectAll("td") .data(table_cells_d3_data_mapper); if (checkboxes.includes("Enable diff view")) { table_cells .filter((d) => { return columns_to_highlight_similarity.includes(d["_key"]); }) .join() .html((cell_data) => { let col_name = cell_data["_key"]; // Original row is the first (yellow highlighted) row in the table let original_value = original_row[col_name]; let comparison_value = cell_data["_orig_data"][col_name]; let diff_view = insert_del_differences( comparison_value, original_value ); return diff_view; }); } // table_cells.style("background-color", (d) => { // // // return potential_match_cells_background_color(d, el); // }); function potential_match_style_each_fn(d) { // Purpose is to call // Ensures we can retrieve this to select const sel = slt.d3.select(this); if (checkboxes.includes("Enable cell shading")) { shade_potential_match_cells(d, sel, el); } } table_cells.each(potential_match_style_each_fn); }3a:T54b,output_data = { const outputs = Object.keys(table_interface).map((key) => { const data = table_interface[key]; if (has_source_dataset_column) { return { [`${source_dataset_column_name}_l`]: original_row[source_dataset_column_name], [`${unique_id_column_name}_l`]: original_row[unique_id_column_name], [`${source_dataset_column_name}_r`]: data[source_dataset_column_name], [`${unique_id_column_name}_r`]: data[unique_id_column_name], match_probability: data.match_probability, match_weight: data.match_weight, clerical_match_score: data.clerical_match_score_value, clerical_notes: data.notes_input_value, original_record_source_dataset: original_row[source_dataset_column_name], original_record_unique_id: original_row[unique_id_column_name] }; } else { return { [`${unique_id_column_name}_l`]: original_row[unique_id_column_name], [`${unique_id_column_name}_r`]: data[unique_id_column_name], match_probability: data.match_probability, match_weight: data.match_weight, clerical_match_score: data.clerical_match_score_value, clerical_notes: data.notes_input_value, original_record_unique_id: original_row[unique_id_column_name] }; } }); return outputs; }3b:T54b,has_existing_label = function (label_pairwise_comparison, comparison_record) { let original_row_id; let comparison_record_id; let pw_comparison_id_l; let pw_comparison_id_r; if (has_source_dataset_column) { original_row_id = original_row[source_dataset_column_name] + "_" + original_row[unique_id_column_name]; comparison_record_id = comparison_record[source_dataset_column_name] + "_" + comparison_record[unique_id_column_name]; pw_comparison_id_l = label_pairwise_comparison[`${source_dataset_column_name}_l`] + "_" + label_pairwise_comparison[`${unique_id_column_name}_l`]; pw_comparison_id_r = label_pairwise_comparison[`${source_dataset_column_name}_r`] + "_" + label_pairwise_comparison[`${unique_id_column_name}_r`]; } else { original_row_id = original_row[unique_id_column_name]; comparison_record_id = comparison_record[unique_id_column_name]; pw_comparison_id_l = label_pairwise_comparison[`${unique_id_column_name}_l`]; pw_comparison_id_r = label_pairwise_comparison[`${unique_id_column_name}_r`]; } return ( (original_row_id == pw_comparison_id_l && comparison_record_id == pw_comparison_id_r) || (original_row_id == pw_comparison_id_r && comparison_record_id == pw_comparison_id_l) ); }3c:Ta91,leven = function (a, b, ratio = true) { // https://github.com/gustf/js-levenshtein/ // MIT License // Copyright (c) 2017 Gustaf Andersson // Permission is hereby granted, free of charge, to any person obtaining a copy // of this software and associated documentation files (the "Software"), to deal // in the Software without restriction, including without limitation the rights // to use, copy, modify, merge, publish, distribute, sublicense, and/or sell // copies of the Software, and to permit persons to whom the Software is // furnished to do so, subject to the following conditions: // The above copyright notice and this permission notice shall be included in all // copies or substantial portions of the Software. a = a.toString(); b = b.toString(); function _min(d0, d1, d2, bx, ay) { return d0 < d1 || d2 < d1 ? d0 > d2 ? d2 + 1 : d0 + 1 : bx === ay ? d1 : d1 + 1; } if (a === b) { if (ratio) { return 1.0; } else { return 0; } } if (a.length > b.length) { var tmp = a; a = b; b = tmp; } var la = a.length; var lb = b.length; while (la > 0 && a.charCodeAt(la - 1) === b.charCodeAt(lb - 1)) { la--; lb--; } var offset = 0; while (offset < la && a.charCodeAt(offset) === b.charCodeAt(offset)) { offset++; } la -= offset; lb -= offset; if (la === 0 || lb < 3) { if (ratio) { return 1 - lb / Math.max(a.length, b.length); } else { return lb; } } var x = 0; var y; var d0; var d1; var d2; var d3; var dd; var dy; var ay; var bx0; var bx1; var bx2; var bx3; var vector = []; for (y = 0; y < la; y++) { vector.push(y + 1); vector.push(a.charCodeAt(offset + y)); } var len = vector.length - 1; for (; x < lb - 3; ) { bx0 = b.charCodeAt(offset + (d0 = x)); bx1 = b.charCodeAt(offset + (d1 = x + 1)); bx2 = b.charCodeAt(offset + (d2 = x + 2)); bx3 = b.charCodeAt(offset + (d3 = x + 3)); dd = x += 4; for (y = 0; y < len; y += 2) { dy = vector[y]; ay = vector[y + 1]; d0 = _min(dy, d0, d1, bx0, ay); d1 = _min(d0, d1, d2, bx1, ay); d2 = _min(d1, d2, d3, bx2, ay); dd = _min(d2, d3, dd, bx3, ay); vector[y] = dd; d3 = d2; d2 = d1; d1 = d0; d0 = dy; } } for (; x < lb; ) { bx0 = b.charCodeAt(offset + (d0 = x)); dd = ++x; for (y = 0; y < len; y += 2) { dy = vector[y]; vector[y] = dd = _min(dy, d0, dd, bx0, vector[y + 1]); d0 = dy; } } if (ratio) { return 1 - dd / Math.max(a.length, b.length); } else { return dd; } }3d:[] 2d:{"title":"Splink Labelling tool","cells":[{"id":0,"value":"# Splink Labelling tool","pinned":false,"hidden":false,"mode":"md"},{"id":2166,"value":"$2f","pinned":false,"hidden":false,"mode":"ojs"},{"id":2170,"value":"$30","pinned":false,"hidden":false,"mode":"ojs"},{"id":1842,"value":"viewof refresh = Inputs.button(\"refresh splink_vis_utils javascript lib\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":1839,"value":"// slt = {\n// if (typeof globalThis.slt == \"undefined\") {\n// let slt = await require(`${localUrl}?${refresh}${Date.now()}`);\n// return slt;\n// } else {\n// return globalThis.slt;\n// }\n// }","pinned":false,"hidden":false,"mode":"ojs"},{"id":1078,"value":"in_settings = null","pinned":false,"hidden":false,"mode":"ojs"},{"id":1981,"value":"blurb = html`\n

Click here for instructions for using this tool

Use this tool to label which records represent a match to the original record, which is highlighted in yellow at the top of the table.

Subsequent rows are possible matches to the original row.

Use the slider on the left hand side to choose a clerical match probability, which is the probability with which you think the possible match is the same entity as the original record.

You may also enter notes to explain why you made a particular decision.

When you're finished labelling, click the download button below the table to export your labels.

\n`","pinned":false,"hidden":false,"mode":"ojs"},{"id":3075,"value":"beta_header = html`\n

BETA\n The Splink labelling tool is still in development, which means some features may change and there may be bugs. Your feedback will help us improve it.

\n`","pinned":false,"hidden":false,"mode":"ojs"},{"id":2010,"value":"$31","pinned":false,"hidden":false,"mode":"ojs"},{"id":2239,"value":"$32","pinned":false,"hidden":false,"mode":"ojs"},{"id":2275,"value":"$33","pinned":false,"hidden":false,"mode":"ojs"},{"id":2294,"value":"draw_potential_matches_spacer_row = function (table) {\n let pm_data = [[\"👇 Potential matches\", \"\", \"\"]];\n const potential_matches_spacer_row = table\n .selectAll(\"tr.potential_matches_spacer_row\")\n .data(pm_data)\n .join(\"tr\")\n .attr(\"class\", \"potential_matches_spacer_row\");\n\n potential_matches_spacer_row\n .selectAll(\"td\")\n .data((d) => d)\n .join(\"td\")\n .text((d) => d)\n .attr(\"class\", (d, i) => {\n if (i == 0) {\n return \"potential_matches_text_with_arrow_cell\";\n } else {\n return \"\";\n }\n });\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2318,"value":"$34","pinned":false,"hidden":false,"mode":"ojs"},{"id":2466,"value":"colspan_spec = {\n let spec = [\n {\n span: 2,\n text: \"User input\",\n css_class: \"user_input_column\"\n }\n ];\n\n let splink_predictions = show_splink_predictions_in_interface.includes(\n \"Show splink predictions\"\n );\n\n if (splink_predictions) {\n spec.push({\n span: 3,\n text: \"Splink predictions\",\n css_class: \"splink_predictions_column\"\n });\n }\n\n spec.push({\n span: 100,\n text: \"Source data\",\n css_class: \"source_data_column\"\n });\n return spec;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2472,"value":"styling_lookup = {\n let lookup = {};\n\n let cumulativeCount = 0;\n\n colspan_spec.forEach((spec) => {\n for (let i = 0; i < spec.span; i++) {\n lookup[cumulativeCount] = {\n css_class: spec.css_class\n };\n cumulativeCount++;\n }\n });\n return lookup;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2248,"value":"$35","pinned":false,"hidden":false,"mode":"ojs"},{"id":2768,"value":"viewof show_splink_predictions_in_interface = {\n let cb = slt.checkbox([\"Show splink predictions\"], {\n value: show_splink_predictions_initial_value\n });\n cb.classList.add(\"labelling-tool-checkbox\");\n return cb;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2945,"value":"$36","pinned":false,"hidden":false,"mode":"ojs"},{"id":2352,"value":"viewof checkboxes = {\n let cb = slt.checkbox([\"Enable cell shading\", \"Enable diff view\"], {\n value: [\"Enable cell shading\"]\n });\n\n cb.classList.add(\"labelling-tool-checkbox\");\n return cb;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":11,"value":"$37","pinned":false,"hidden":false,"mode":"ojs"},{"id":1792,"value":"dl = {\n let button = DOM.download(\n new Blob([slt.d3.csvFormat(output_data)], { type: \"text/csv\" }),\n output_filename,\n \"⬇️ Download labels in .csv format\"\n );\n\n button.classList.add(\"labelling-tool-checkbox\");\n return button;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2913,"value":"copy_to_clipboard_button = {\n const button = html``;\n button.onclick = () => {\n const csvData = slt.d3.csvFormat(output_data);\n window.navigator.clipboard.writeText(csvData);\n };\n\n button.classList.add(\"labelling-tool-checkbox\");\n\n return button;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":3246,"value":"labels_in_textarea = html`\n

Note for JupyterLab users

If you're viewing this within the Jupyter Lab html viewer, html is sandboxed so the 'download labels' button will not work.

As a workaround, you can manually copy and paste the following text and use\npd.read_clipboard() to load this data into a pandas dataframe in Python.

\n${slt.d3.csvFormat(output_data)}\n\n`","pinned":false,"hidden":false,"mode":"ojs"},{"id":1640,"value":"tooltip = slt.d3\n  .select(\"body\")\n  .append(\"div\")\n  .attr(\"class\", \"d3-tooltip\")\n  .style(\"z-index\", \"10\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":1637,"value":"### Styling","pinned":false,"hidden":false,"mode":"md"},{"id":763,"value":"$38","pinned":false,"hidden":false,"mode":"ojs"},{"id":534,"value":"$39","pinned":false,"hidden":false,"mode":"ojs"},{"id":463,"value":"table_cells_d3_data_mapper = function (row_data) {\n  // Transform the row into an array of values\n  // one for each cell.\n  // The values are themselves dictionaries, so that the\n  // cell can be aware of not only its value, but also e.g. the row id\n  let cell_data = Object.entries(row_data);\n\n  let uid_col = \"unique_id\";\n\n  cell_data = cell_data.map((entry) => {\n    let key = entry[0];\n    let value = entry[1];\n    let cell_datum = {};\n    cell_datum[\"_key\"] = key;\n    cell_datum[\"_td_contents\"] = value;\n    cell_datum[\"_row_uid\"] = row_data[uid_col];\n    cell_datum[\"_orig_data\"] = row_data;\n\n    return cell_datum;\n  });\n\n  cell_data = cell_data.filter((d) => d[\"_key\"] != \"_row_num\");\n\n  cell_data = cell_data.filter((entry) => splink_scores_filter(entry));\n\n  return cell_data;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":434,"value":"form_oninput_update_value = function (original_cell_data, el) {\n  // When we see an oninput event no any of the form elements (notes input, range input)\n  // update the value exposed by the form\n\n  let uid = original_cell_data[\"_row_uid\"];\n\n  if (original_cell_data[\"_key\"] == \"clerical_match_score\") {\n    el.value[uid][\"clerical_match_score_value\"] =\n      original_cell_data[\"_td_contents\"].value;\n  }\n\n  if (original_cell_data[\"_key\"] == \"notes_input\") {\n    el.value[uid][\"notes_input_value\"] =\n      original_cell_data[\"_td_contents\"].value;\n  }\n\n  el.dispatchEvent(new Event(\"input\", { bubbles: true }));\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1288,"value":"## Dataflow","pinned":false,"hidden":false,"mode":"md"},{"id":1076,"value":"splink_settings = {\n  let ss;\n  if (typeof globalThis.splink_settings_from_splink == \"undefined\") {\n    ss = default_splink_settings;\n  } else {\n    ss = globalThis.splink_settings_from_splink;\n  }\n  return ss;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1065,"value":"\npairwise_comparison_data = {\n  let pcd;\n  if (typeof globalThis.pairwise_comparison_data_from_splink == \"undefined\") {\n    pcd = default_pairwise_comparison_data;\n  } else {\n    pcd = globalThis.pairwise_comparison_data_from_splink;\n  }\n\n\n  return pcd;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1212,"value":"pairwise_comparison_data_formatted = {\n  let formatted_data = pairwise_comparison_data.map((d) => {\n    return Object.assign({}, d);\n  });\n\n  formatted_data.map((d) => {\n    d[\"match_weight\"] = match_prob_number_format(d[\"match_weight\"]);\n  });\n\n  formatted_data.map((d) => {\n    d[\"match_probability\"] = match_prob_number_format(d[\"match_probability\"]);\n  });\n\n  formatted_data = formatted_data.filter((d) => !is_self_match(d));\n\n  if (splink_settings.link_type == \"link_only\") {\n    formatted_data = formatted_data.filter((d) => !has_same_source_dataset(d));\n  }\n\n\n  return formatted_data;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":235,"value":"original_row = {\n  return get_original_record(\n    pairwise_comparison_data,\n    original_record_suffix,\n    unique_id_column_name\n  );\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":5,"value":"potential_matches_data = {\n  return get_potential_matches(\n    pairwise_comparison_data_formatted,\n    comparison_records_suffix,\n    unique_id_column_name\n  );\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":133,"value":"data_with_inputs = {\n  let new_data = potential_matches_data.map((pairwise_comparison, i) => {\n    let match_score = 0.0;\n    let notes_value = \"\";\n\n    if (existing_labels_csv != null) {\n      // Check if thre is a row in existing_labels_csv\n      // which matches d[source_dataest_]\n\n      existing_labels_csv.forEach((existing_label) => {\n        if (has_existing_label(existing_label, pairwise_comparison)) {\n          match_score = existing_label[\"clerical_match_score\"];\n          notes_value = existing_label[\"clerical_notes\"];\n        }\n      });\n    }\n\n    // Look up if exists in existing labels\n    var clone = Object.assign(\n      {\n        clerical_match_score: slt.range([0, 1], {\n          value: match_score,\n          step: 0.001\n        }),\n        notes_input: slt.textarea({ value: notes_value })\n      },\n      pairwise_comparison\n    );\n    clone[\"_row_num\"] = i;\n\n    return clone;\n  });\n\n  return new_data;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1775,"value":"$3a","pinned":false,"hidden":false,"mode":"ojs"},{"id":2903,"value":"output_filename = {\n  if (has_source_dataset_column) {\n    return `labels_${original_row[source_dataset_column_name]}_${original_row[unique_id_column_name]}.csv`;\n  } else {\n    return `labels_${original_row[unique_id_column_name]}.csv`;\n  }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2745,"value":"## Constants","pinned":false,"hidden":false,"mode":"md"},{"id":1164,"value":"comparison_records_suffix = get_suffix_of_comparison_records_l_or_r(\n  pairwise_comparison_data,\n  unique_id_column_name\n)","pinned":false,"hidden":false,"mode":"ojs"},{"id":1151,"value":"original_record_suffix = get_suffix_of_original_record_l_or_r(\n  pairwise_comparison_data,\n  unique_id_column_name\n)","pinned":false,"hidden":false,"mode":"ojs"},{"id":1088,"value":"unique_id_column_name = {\n  return splink_settings[\"unique_id_column_name\"] ?? \"unique_id\";\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1101,"value":"source_dataset_column_name = {\n  return splink_settings[\"source_dataset_column_name\"] ?? \"source_dataset\";\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":524,"value":"columns_to_highlight_similarity = {\n  // Get the co\n\n  let remove_columns = new Set([\n    unique_id_column_name,\n    source_dataset_column_name\n  ]);\n  additional_columns_to_retain.forEach((d) => remove_columns.add(d));\n\n  let all_keys = Object.keys(pairwise_comparison_data[0]);\n\n  all_keys = all_keys.filter((d) => d.endsWith(\"_l\"));\n  all_keys = all_keys.filter((d) => !d.startsWith(\"tf_\"));\n  all_keys = all_keys.map((d) => d.slice(0, -2));\n\n  return all_keys.filter((value) => {\n    return !remove_columns.has(value);\n  });\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1097,"value":"additional_columns_to_retain = {\n  return splink_settings[\"additional_columns_to_retain\"] ?? [];\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":3024,"value":"has_source_dataset_column = {\n  let example_row = pairwise_comparison_data[0];\n\n  return `${source_dataset_column_name}_l` in example_row;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1239,"value":"## Pure functions","pinned":false,"hidden":false,"mode":"md"},{"id":2616,"value":"is_self_match = function (record) {\n  // Since splink uses `linker.find_matches_to_new_records` under the hood, and this method\n  // assumes it's a new, not an existing record, we need to filter out the 'self match'\n\n  if (has_source_dataset_column) {\n    // Filter out records with both source_dataset_column_name and unique_id_column_name equal to those in the example_row\n    return (\n      record[`${source_dataset_column_name}_l`] ==\n        record[`${source_dataset_column_name}_r`] &&\n      record[`${unique_id_column_name}_l`] ==\n        record[`${unique_id_column_name}_r`]\n    );\n  } else {\n    // Filter out records with unique_id_column_name equal to that in the example_row\n    return (\n      record[`${unique_id_column_name}_l`] ==\n      record[`${unique_id_column_name}_r`]\n    );\n  }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2755,"value":"has_same_source_dataset = function (record) {\n  if (has_source_dataset_column) {\n    return (\n      record[`${source_dataset_column_name}_l`] ==\n      record[`${source_dataset_column_name}_r`]\n    );\n  } else {\n    return true;\n  }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1242,"value":"get_output_columns = function (find_matches_data) {\n  // Get columns - do we assume we have settings?\n  // Find columns ending in `_l`, this is the list of columns\n  let first_row = find_matches_data[0];\n\n  const output_columns = Object.keys(first_row)\n    .filter((d) => d.endsWith(\"_l\"))\n    .filter((d) => !d.startsWith(\"tf_\"))\n    .map((d) => d.slice(0, -2));\n\n  output_columns.unshift(\"match_probability\");\n  output_columns.unshift(\"match_weight\");\n  return output_columns;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2990,"value":"$3b","pinned":false,"hidden":false,"mode":"ojs"},{"id":2789,"value":"splink_scores_filter = function (data) {\n  let show_splink_scores = show_splink_predictions_in_interface.includes(\n    \"Show splink predictions\"\n  );\n\n  if (show_splink_scores) {\n    return true;\n  } else {\n    return ![\n      \"match_probability\",\n      \"match_weight\",\n      \"model_truth_status\"\n    ].includes(data._key);\n  }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1245,"value":"get_suffix_of_original_record_l_or_r = function (\n  find_matches_data,\n  unique_id_colname\n) {\n  let left_ids = new Set();\n  let right_ids = new Set();\n\n  find_matches_data.forEach((d) => left_ids.add(d[unique_id_colname]));\n  find_matches_data.forEach((d) => right_ids.add(d[unique_id_colname]));\n\n  // Preference for right because I believe in current version of Splink (3.4.2)\n  // you're guaranteed the original is on the right\n\n  if (right_ids.size == 1) {\n    return \"_r\";\n  }\n\n  if (left_ids.size == 1) {\n    return \"_l\";\n  }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1251,"value":"get_suffix_of_comparison_records_l_or_r = function (\n  find_matches_data,\n  unique_id_colname\n) {\n  let orig_suffic = get_suffix_of_original_record_l_or_r(\n    find_matches_data,\n    unique_id_colname\n  );\n  let comp_suffix = \"_l\";\n  if (orig_suffic == \"_l\") {\n    comp_suffix = \"_r\";\n  }\n  return comp_suffix;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1254,"value":"transform_pairwise_comparison_to_label_format = function (\n  pairwise_comparison,\n  suffix,\n  unique_id_colname,\n  original_record\n) {\n  let transformed_record = {};\n\n  transformed_record[\"match_probability\"] =\n    pairwise_comparison[\"match_probability\"];\n  transformed_record[\"match_weight\"] = pairwise_comparison[\"match_weight\"];\n  transformed_record[\"model_truth_status\"] = null;\n  if (original_record) {\n    transformed_record[\"match_probability\"] = null;\n    transformed_record[\"match_weight\"] = null;\n  }\n  transformed_record[unique_id_colname] =\n    pairwise_comparison[`${unique_id_colname}${suffix}`];\n\n  Object.keys(pairwise_comparison).forEach((key) => {\n    if (key.endsWith(suffix) & !key.startsWith(\"tf_\")) {\n      let key_no_suffix = key.slice(0, -2);\n      transformed_record[key_no_suffix] = pairwise_comparison[key];\n    }\n  });\n\n  return transformed_record;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1256,"value":"get_original_record = function (find_matches_data, suffix, unique_id_colname) {\n  \n  let first_record = find_matches_data.filter((d) => is_self_match(d))[0];\n  return transform_pairwise_comparison_to_label_format(\n    first_record,\n    suffix,\n    unique_id_colname\n  );\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1258,"value":"get_potential_matches = function (\n  find_matches_data,\n  suffix,\n  unique_id_colname\n) {\n  let label_format_data = find_matches_data.map((d) => {\n    return transform_pairwise_comparison_to_label_format(\n      d,\n      suffix,\n      unique_id_colname\n    );\n  });\n\n  let sorted_data = label_format_data.sort((a, b) => {\n    return parseFloat(b.match_weight) - parseFloat(a.match_weight);\n  });\n\n  return sorted_data;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":611,"value":"$3c","pinned":false,"hidden":false,"mode":"ojs"},{"id":2221,"value":"jaro_winkler_distance = function (a, b) {\n  var a = a.toString();\n  var b = b.toString();\n  return slt.jw_distance(a, b);\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":2199,"value":"function insert_del_differences(originalValue, comparisonValue) {\n  let diffs = slt.diffChars(comparisonValue, originalValue);\n\n  function get_span(p) {\n    const bgc = p.added ? \"#eaf2c2\" : p.removed ? \"#fadad7\" : \"white\";\n    const textc = p.added ? \"#406619\" : p.removed ? \"#b30000\" : \"black\";\n\n    return `<span style=\"background: ${bgc}; color: ${textc}\">${p.value}</span>`;\n  }\n\n  return `${diffs.map((part) => get_span(part)).join(\"\")}`;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":3204,"value":"function score_to_truth_status(clerical_score, splink_score) {\n  if (clerical_score > 0.5 && splink_score > 0.5) {\n    return \"TP\";\n  } else if (clerical_score <= 0.5 && splink_score <= 0.5) {\n    return \"TN\";\n  } else if (clerical_score > 0.5 && splink_score <= 0.5) {\n    return \"FN\";\n  } else {\n    return \"FP\";\n  }\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":1305,"value":"### Formatting","pinned":false,"hidden":false,"mode":"md"},{"id":1260,"value":"match_prob_number_format = function (num) {\n  return num.toFixed(4);\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":375,"value":"red_orange_green_colour_scale = slt.d3\n  .scaleLinear()\n  .domain([0, 0.5, 1])\n  .range([\"red\", \"orange\", \"green\"])","pinned":false,"hidden":false,"mode":"ojs"},{"id":3146,"value":"red_white_green_colour_scale = slt.d3\n  .scaleLinear()\n  .domain([0, 0.1, 0.5, 0.9, 1])\n  .range([\"#FF9898\", \"#FFDBDB\", \"white\", \"#E5FFE5\", \"limegreen\"])","pinned":false,"hidden":false,"mode":"ojs"},{"id":648,"value":"white_green_colour_scale = slt.d3\n  .scaleLinear()\n  .domain([0, 0.7, 1])\n  .range([\"white\", \"rgb(214, 245, 214)\", \"limegreen\"])","pinned":false,"hidden":false,"mode":"ojs"},{"id":3155,"value":"white_green_colour_scale_match_prob = slt.d3\n  .scaleLinear()\n  .domain([0, 1])\n  .range([\"white\", \"limegreen\"])","pinned":false,"hidden":false,"mode":"ojs"},{"id":735,"value":"white_blue_colour_scale = slt.d3\n  .scaleLinear()\n  .domain([0, 0.75, 1])\n  .range([\"white\", \"rgb(143, 200, 255)\", \"dodgerblue\"])","pinned":false,"hidden":false,"mode":"ojs"},{"id":1557,"value":"## Constants","pinned":false,"hidden":false,"mode":"md"},{"id":1561,"value":"table_cell_padding = \"5px\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":1563,"value":"transitive_highlight_width = \"4px\"","pinned":false,"hidden":false,"mode":"ojs"},{"id":1573,"value":"padding_with_highlight = {\n  return (\n    parseInt(table_cell_padding) - parseInt(transitive_highlight_width) + \"px\"\n  );\n}\n","pinned":false,"hidden":false,"mode":"ojs"},{"id":3273,"value":"show_splink_predictions_initial_value = {\n  if (typeof globalThis.SHOW_SPLINK_PREDICTIONS == \"undefined\") {\n    return [\"Show splink predictions\"];\n  } else if (globalThis.SHOW_SPLINK_PREDICTIONS == true) {\n    return [\"Show splink predictions\"];\n  } else {\n    return [];\n  }\n}","pinned":true,"hidden":false,"mode":"ojs"},{"id":1834,"value":"## Embedding things\n","pinned":false,"hidden":false,"mode":"md"},{"id":1836,"value":"localUrl = \"http://127.0.0.1:8080/dist/slt.js\"","pinned":true,"hidden":false,"mode":"ojs"},{"id":343,"value":"## TODO:\n\n\n- The nested join `.data()` call for the table cells (`td`) is where we should compute the direct and transitive matches and provide data to the cells accordingly.  e.g. data could be a dict like `{value:, direct_matchiness: 0.8, indirect_matchiness: 0.7}`\n","pinned":false,"hidden":false,"mode":"md"}],"files":"$W3d","stdlib":"1","titling":"auto"}
</article></body></html>