Markup case names #241

flooie · 2025-04-02T17:39:45Z

This PR adds support for identifying the full case name in markup text. When a plaintiff is identified
it calls a new method to check if the text is inside a tag and extracts that out.

Also supports antecedent guess

This should improve reference citation extraction and better identify citations.

Add check inside method for plaintiff Use that to identify the full plaintiff name Also improve regex building to avoid unnecessary escapes

Identify antecedent guesses in markup change method name

grossir

looking good, some comments below

eyecite/helpers.py

Changed method for speed improvements Added improvements for defendants as well. More example tests Fixed full span

flooie · 2025-04-03T16:19:13Z

---

Co.                                                                    --->      Silverthorne Lumber Co.
Silverthorne Lumber Co. v. United States, 251 U. S. 385
---

Portelli, 15 N. Y. 2d 235                                              --->      Portelli
People v. Portelli, 15 N. Y. 2d 235, 205 N. E. 2d 857
---

States                                                                 --->      United States
United States v. Grunewald, 233 F. 2d 556
---

States                                                                 --->      United States
Grunewald, 233 F. 2d 556, 579, 581-582 (Frank, J., dissenting), rev'd  --->      Grunewald
United States v. Grunewald, 233 F. 2d 556, 579, 581-582 (Frank, J., dissenting), rev'd, 353 U. S. 391
---

Dorado, 62 Cal. 2d 338, 351                                            --->      Dorado
People v. Dorado, 62 Cal. 2d 338, 351, 398 P. 2d 361
---

Dorado, 62 Cal. 2d 338, 351, 398 P. 2d 361, 369-370                    --->      Dorado
People v. Dorado, 62 Cal. 2d 338, 351, 398 P. 2d 361, 369-370, 42 Cal. Rptr. 169
---

U. S., 337 F. 2d 136 (1964), cert. den.                                --->      U. S.
Jackson v. U. S., 337 F. 2d 136 (1964), cert. den. 380 U. S. 935
---

S.                                                                     --->      U. S.
U. S. v. Konigsberg, 336 F. 2d 844
---

S.                                                                     --->      U. S.
Konigsberg, 336 F. 2d 844 (1964), cert. den.                           --->      Konigsberg
U. S. v. Konigsberg, 336 F. 2d 844 (1964), cert. den. 379 U. S. 933
---

States                                                                 --->      United States
United States v. Scully, 225 F. 2d 113
---

States                                                                 --->      United States
Carignan, *528                                                         --->      Carignan
United States v. Carignan, *528 342 U. S. 36
---

Hansen                                                                 --->      Sparf and Hansen
Sparf and Hansen v. United States, 156 U. S. 51
---

States                                                                 --->      United States
United States v. Mitchell, 322 U. S. 65
---

States                                                                 --->      United States
United States v. Carignan, 342 U. S. 36
---

States                                                                 --->      United States
United States v. Romano, 382 U. S. 136
---

State, 238 Md. 140                                                     --->      State
Johnson v. State, 238 Md. 140, 207 A. 2d 643
---

Modesto, 62 Cal. 2d 436, 446                                           --->      Modesto
People v. Modesto, 62 Cal. 2d 436, 446, 398 P. 2d 753
---

States                                                                 --->      United States
United States v. Drummond, 354 F. 2d 132
---

States                                                                 --->      United States
United States v. Childress, 347 F. 2d 448
---

Dorado, 62 Cal. 2d 338                                                 --->      Dorado
People v. Dorado, 62 Cal. 2d 338, 398 P. 2d 361
---

Dorado, 62 Cal. 2d 338, 398 P. 2d 361                                  --->      Dorado
People v. Dorado, 62 Cal. 2d 338, 398 P. 2d 361, 42 Cal. Rptr. 169
---

Hartgraves, 31 Ill. 2d 375                                             --->      Hartgraves
People v. Hartgraves, 31 Ill. 2d 375, 202 N. E. 2d 33
---

Wakat, 415 Ill. 610                                                    --->      Wakat
People v. Wakat, 415 Ill. 610, 114 N. E. 2d 706
---

State, 213 Md. 556                                                     --->      State
Kier v. State, 213 Md. 556, 132 A. 2d 494
---

People, 113 Colo. 194                                                  --->      People
Bruner v. People, 113 Colo. 194, 156 P. 2d 111
---

Matlock, 51 Cal. 2d 682                                                --->      Matlock
People v. Matlock, 51 Cal. 2d 682, 336 P. 2d 505
---

States                                                                 --->      United States
United States v. Murphy, 222 F. 2d 698
---

Bonino, 1 N. Y. 2d 752                                                 --->      Bonino
People v. Bonino, 1 N. Y. 2d 752, 135 N. E. 2d 51
---

States                                                                 --->      United States
United States v. Carignan, 342 U. S. 36
---

Donovan, 13 N. Y. 2d 148                                               --->      Donovan
People v. Donovan, 13 N. Y. 2d 148, 193 N. E. 2d 628
---

Brown                                                                  --->      United States ex rel. Brown
United States ex rel. Brown v. Fay, 242 F. Supp. 273
---

Witenski, 15 N. Y. 2d 392                                              --->      Witenski
People v. Witenski, 15 N. Y. 2d 392, 207 N. E. 2d 358
---

Dorado, 62 Cal. 2d 338, 354                                            --->      Dorado
People v. Dorado, 62 Cal. 2d 338, 354, 398 P. 2d 361
---

Dorado, 62 Cal. 2d 338, 354, 398 P. 2d 361, 371                        --->      Dorado
People v. Dorado, 62 Cal. 2d 338, 354, 398 P. 2d 361, 371, 42 Cal. Rptr. 169
---

States                                                                 --->      United States
United States v. Rose, 24 CMR 251
---

States                                                                 --->      United States
United States v. Gunnels, 23 CMR 354
---

Angelet                                                                --->      United States ex rel. Angelet
United States ex rel. Angelet v. Fay, 333 F. 2d 12
---

Inc.                                                                   --->      Ziffrin, Inc.
Ziffrin, Inc. v. United States, 318 U. S. 73
---

Morse, 60 Cal. 2d 631                                                  --->      Morse
People v. Morse, 60 Cal. 2d 631, 388 P. 2d 33
---

Morse, 60 Cal. 2d 631, 388 P. 2d 33                                    --->      Morse
People v. Morse, 60 Cal. 2d 631, 388 P. 2d 33, 36 Cal. Rptr. 201
---

States                                                                 --->      United States
United States v. Carignan, 342 U. S. 36
---

Co.                                                                    --->      Shotwell Mfg. Co.
Shotwell Mfg. Co. v. United States, 371 U. S. 341
---

States                                                                 --->      United States
United States v. Robinson, 354 F. 2d 109
---

Townsend                                                               --->      United States ex rel. Townsend
United States ex rel. Townsend v. Ogilvie, 334 F. 2d 837
---

Hartgraves, 31 Ill. 2d 375                                             --->      Hartgraves
People v. Hartgraves, 31 Ill. 2d 375, 202 N. E. 2d 33
---

Fox, ___ Iowa ___                                                      --->      Fox
State v. Fox, ___ Iowa ___, 131 N. W. 2d 684
---

Warden, 236 Md. 236                                                    --->      Warden
Parker v. Warden, 236 Md. 236, 203 A. 2d 418
---

State, ___ Nev. ___                                                    --->      State
Bean v. State, ___ Nev. ___, 398 P. 2d 251
---

Hodgson, 44 N. J. 151                                                  --->      Hodgson
State v. Hodgson, 44 N. J. 151, 207 A. 2d 542
---

Gunner, 15 N. Y. 2d 226                                                --->      Gunner
People v. Gunner, 15 N. Y. 2d 226, 205 N. E. 2d 852
---

Linde                                                                  --->      Commonwealth ex rel. Linde
Commonwealth ex rel. Linde v. Maroney, 416 Pa. 331
---

Linde                                                                  --->      Commonwealth ex rel. Linde
Maroney, 416 Pa. 331                                                   --->      Maroney
Commonwealth ex rel. Linde v. Maroney, 416 Pa. 331, 206 A. 2d 288
---

State, 24 Wis. 2d 491                                                  --->      State
Browne v. State, 24 Wis. 2d 491, 131 N. W. 2d 169
---

Russo                                                                  --->      United States ex rel. Russo
United States ex rel. Russo v. New Jersey, 351 F. 2d 429
---

Dorado, 62 Cal. 2d 338                                                 --->      Dorado
People v. Dorado, 62 Cal. 2d 338, 398 P. 2d 361
---

Dufour, ___ R. I. ___                                                  --->      Dufour
State v. Dufour, ___ R. I. ___, 206 A. 2d 82
---

Neely, 239 Ore. 487                                                    --->      Neely
State v. Neely, 239 Ore. 487, 395 P. 2d 557
---

Neely, 239 Ore. 487, 395 P. 2d 557, modified                           --->      Neely
State v. Neely, 239 Ore. 487, 395 P. 2d 557, modified, 398 P. 2d 482
---

Dorado, 62 Cal. 2d 338                                                 --->      Dorado
People v. Dorado, 62 Cal. 2d 338, 398 P. 2d 361
---

States                                                                 --->      United States
United States v. Bolden, 355 F. 2d 453
---

Du Bont, 235 Cal. App. 2d 844                                          --->      Du Bont

I ran a test on Miranda v. Arizona to illustrate the improvements. It lists only changes identified.

In this case you can see accurate plaintiff parsing and you can also see better defendant parsing. Especially when or usually when the citation is a parallel to the main. It no longer mistakenly identifies the previous citation with the defendant name when handling markup text.

Refactor methods for readability Add support for antecedents Strip out stop words if starting extraction

Pagination has long been in the way and this helps identify cites with pagination in the middle of the citation

Add test to illustrate L.A. support Can now extract defendant out properly

grossir

The names that this PR are able to collect look great.

However, there are some weird losses on the benchmark

Akins

Since it's picking everything inside the emphasis tag, it is picking 'Cf' as part of the name. This makes it lose the following Akins standalone reference, since it's looking for Cf. Akins. Solvable by cleaning up stopwords like Cf. Another one that happens a lot inside the same style tags as the name is "See", you should probably include these in the tests

 FullCaseCitation('104 So.3d 1173', groups={'volume': '104', 'reporter': 'So.3d', 'page': '1173'}, metadata=FullCaseCitation.Metadata(parenthetical='reversing nonhomicide sentence of life without parole where trial court exercised its discretion in imposing a sentence of twenty-seven years on the related homicide offense of second-degree murder, an offense punishable by a maximum of life in prison', pin_cite=None, pin_cite_span_start=None, pin_cite_span_end=None, year='2012', court=None, plaintiff='Cf. Akins', defendant='State', extra=None, antecedent_guess=None, resolved_case_name_short=None, resolved_case_name=None)),

Foy

This is the full citation "The opinion in Thrift Funds Canal, Inc. v. Foy, 261 La. 573, 260 So.2d 628 (1972), classifies conventional mortgages into three categories:"

On this branch, it's not picking up "Foy" as defendant

 FullCaseCitation('261 La. 573', groups={'volume': '261', 'reporter': 'La.', 'page': '573'}, metadata=FullCaseCitation.Metadata(parenthetical=None, pin_cite=None, pin_cite_span_start=None, pin_cite_span_end=None, year='1972', court=None, plaintiff='Inc.', defendant=None, extra='260 So.2d 628', antecedent_guess=None, resolved_case_name_short=None, resolved_case_name=None)),
 FullCaseCitation('260 So.2d 628', groups={'volume': '260', 'reporter': 'So.2d', 'page': '628'}, metadata=FullCaseCitation.Metadata(parenthetical=None, pin_cite=None, pin_cite_span_start=None, pin_cite_span_end=None, year='1972', court=None, plaintiff='Inc.', defendant=None, extra=None, antecedent_guess=None, resolved_case_name_short=None, resolved_case_name=None)),

but on main, it is picking it up. I am not sure about the reason for the change...

 FullCaseCitation('261 La. 573', groups={'volume': '261', 'reporter': 'La.', 'page': '573'}, metadata=FullCaseCitation.Metadata(parenthetical=None, pin_cite=None, pin_cite_span_start=None, pin_cite_span_end=None, year='1972', court=None, plaintiff='Inc.', defendant='Foy', extra='260 So.2d 628', antecedent_guess=None, resolved_case_name_short=None, resolved_case_name=None)),
 FullCaseCitation('260 So.2d 628', groups={'volume': '260', 'reporter': 'So.2d', 'page': '628'}, metadata=FullCaseCitation.Metadata(parenthetical=None, pin_cite=None, pin_cite_span_start=None, pin_cite_span_end=None, year='1972', court=None, plaintiff='Inc.', defendant='Foy', extra=None, antecedent_guess=None, resolved_case_name_short=None, resolved_case_name=None)),

Which makes it able to pick up 2 references below

You should probably check the "gain" column (which is actually the loss)

eyecite/helpers.py

grossir · 2025-04-03T19:30:32Z

eyecite/helpers.py

+                if (
+                    document.markup_text
+                    and document.plain_to_markup
+                    and document.markup_to_plain
+                ):


these checks are also repeated in lines 203-207, shouldn't they just be done inside the functions themselves? And just return the same input values if there are no markup related attributes

In fact, a couple of the checks are already repeated inside of extract_full_text_from_markup, but not inside the other one

# Convert plain text offset to a markup offset. and bail if no markup if not document.plain_to_markup or not document.markup_to_plain: return base_offset, default_text

this was done to satisfy mypy

maybe it's a good time to use the mypy ignore
https://mypy.readthedocs.io/en/stable/common_issues.html#spurious-errors-and-locally-silencing-the-checker

eyecite/helpers.py

grossir · 2025-04-03T19:32:13Z

eyecite/helpers.py

+                new_text = update_defendant_markup(document, stop_word)
+                if citation.metadata.defendant != new_text:
+                    citation.metadata.defendant = new_text


if the text is updated, shouldn't the full_span_start be updated too?

this wont affect the start, this actually just reduces the defendant on the back end. Thats why no update.

grossir · 2025-04-03T19:33:04Z

eyecite/helpers.py

+    Parameters:
+        document: The Document instance.
+        base_offset: The plain text offset to convert
+        default_text: The current text stored for this field.
+        source_end: Optional plain text end index to use when extracting text.
+                    If provided, full text is extracted as:
+                        document.plain_text[plain_text_start:source_end]


I think we should keep the standard docstring format we are using in eyecite

:param document: ... :param base_offset: ...

actually I think the format is somewhere in the middle

(e.g., a judicial opinion or other legal doc), return a list of `eyecite.models.CitationBase` objects representing the citations found in the doc. Args: plain_text: The text to parse. You may wish to use the `eyecite.clean.clean_text` function to pre-process your text before passing it here. remove_ambiguous: Whether to remove citations that might refer to more than one reporter and can't be narrowed down by date. tokenizer: An instance of a Tokenizer object. See `eyecite.tokenizers` for information about available tokenizers. Uses the `eyecite.tokenizers.AhocorasickTokenizer` by default. markup_text: if the source text has markup (XML or HTML mostly), pass it to extract ReferenceCitations that may be detectable via markup style tags clean_steps: Cleanup steps and methods Returns: A list of `eyecite.models.CitationBase` objects

grossir · 2025-04-03T19:34:40Z

eyecite/helpers.py

+        # Not inside an emphasis/i tag
+        return base_offset, default_text
+
+    # Use the first and only matching emphasis tag.


there are no checks in place for to ensure it is the "only" tag in filtered_results

grossir · 2025-04-03T19:35:27Z

eyecite/helpers.py

+        document: The Document instance.
+        base_offset: The plain text offset to convert
+        default_text: The current text stored for this field.
+        source_end: Optional plain text end index to use when extracting text.


why not call this argument plain_text_end instead of source_end

flooie · 2025-04-03T20:23:01Z

Judge Regan (dissenting) in <i>Thrift Funds Canal,</i> Inc. v. Foy, 242 So.2d 253, 257 (La.App. 4 Cir. 1970), calls

we lose FOY because its not correctly tagged in HTML

Add Cf. as stop word Dont ovewrite none if we have data Update docstrings and comments

More edge case examples

grossir

This looks great, I think it's mostly ready

My only nitpick is that add_defendant is now a big and complex function, and adds both plaintiff and defendant, and we should refactor it eventually

grossir · 2025-04-04T01:10:20Z

eyecite/helpers.py

+                if (
+                    document.markup_text
+                    and document.plain_to_markup
+                    and document.markup_to_plain
+                ):


maybe it's a good time to use the mypy ignore
https://mypy.readthedocs.io/en/stable/common_issues.html#spurious-errors-and-locally-silencing-the-checker

mlissner · 2025-04-04T05:35:07Z

My only nitpick is that add_defendant is now a big and complex function, and adds both plaintiff and defendant, and we should refactor it eventually

I haven't looked at the PR, but I think now may be the best time to clean things up, since we probably won't return to eyecite for a while, and when we do it'll be harder to do the refactoring.

Use type ignore and flake8

flooie · 2025-04-04T16:05:44Z

My only nitpick is that add_defendant is now a big and complex function, and adds both plaintiff and defendant, and we should refactor it eventually

I haven't looked at the PR, but I think now may be the best time to clean things up, since we probably won't return to eyecite for a while, and when we do it'll be harder to do the refactoring.

Just seeing this. It did both plaintiff and defendant previously. But plaintiff was always a throw-a-way process.

This is a large refactor It is meant to improve case title extraction by removing add defendants and relying on either HTML or a series of heurestics to make educated guesses about when a title stops It separates the logic between html and plaintext It adds to stop words it removes add defendant

I didnt update a test for our new improved extraction. This one caputres the full text

flooie · 2025-04-05T00:09:33Z

@grossir something is cached or broken about the benchmark ... I found one in particular weird but others I couldn't replicate.

this entry ```

1744543		I am not going to let you represent yourself


made no sense - so I ran it locally to test it out 

[
    {
        "cites": [
            "\u00a7",
            "\u00a7",
            "\u00a7",
            "422 U.S. 806",
            "95 S.Ct. 2525",
            "45 L.Ed.2d 562",
            "\u00a7",
            "542 S.W.2d 549",
            "685 S.W.2d 942",
            "587 S.W.2d 611",
            "589 F.2d 45",
            "77 Cal.App.3d 45",
            "143 Cal.Rptr. 379",
            "270 Ind. 55",
            "383 N.E.2d 309",
            "267 Ind. 289",
            "370 N.E.2d 318",
            "95 Nev. 933",
            "604 P.2d 338",
            "590 P.2d 679",
            "588 P.2d 572",
            "98 A.L.R.3d 13",
            "\u00a7",
            "572 S.W.2d 440",
            "622 S.W.2d 416",
            "\u00a7",
            "\u00a7",
            "\u00a7",
            "\u00a7",
            "647 S.W.2d 799",
            "\u00a7",
            "558 S.W.2d 646",
            "593 S.W.2d 208",
            "\u00a7",
            "694 S.W.2d 903",
            "603 S.W.2d 569",
            "676 S.W.2d 304",
            "664 S.W.2d 617",
            "463 U.S. 277",
            "103 S.Ct. 3001",
            "77 L.Ed.2d 637",
            "supra,",
            "\u00a7\u00a7",
            "\u00a7\u00a7",
            "\u00a7",
            "\u00a7",
            "\u00a7\u00a7",
            "\u00a7\u00a7",
            "\u00a7\u00a7",
            "\u00a7\u00a7",
            "\u00a7\u00a7",
            "\u00a7\u00a7",
            "\u00a7\u00a7",
            "\u00a7",
            "\u00a7",
            "\u00a7\u00a7",
            "\u00a7",
            "\u00a7",
            "\u00a7",
            "459 U.S. 359",
            "103 S.Ct. 673",
            "74 L.Ed.2d 535",
            "647 S.W.2d 799",
            "624 S.W.2d 16",
            "689 S.W.2d 758",
            "657 S.W.2d 677",
            "653 S.W.2d 227",
            "supra,",
            "664 S.W.2d 617",
            "603 S.W.2d 683",
            "588 S.W.2d 65"
        ],
        "time": 0.243268
    }
]

and this was the output - you can see there is no random long string.

flooie · 2025-04-05T00:31:19Z

@grossir there we go - I need to re run name extraction because many html files have some tagged nicely and some not.

Fix bug in index shift and text cleanup

github-actions · 2025-04-05T01:01:34Z

The Eyecite Report 👁️

Gains and Losses

There were 5 gains and 21 losses.

Click here to see details.

There were 53 changes so we are only displaying the first 50. You can review the
entire list by downloading the output.csv file linked above.

id	Gain	Loss
5329531		German Savings Bank
5656104	Sime
6207147	Arrington
1917661		Vanner
1662392		Belton
1783747		Parish of East Baton Rouge
1537257		St. Cloud
1717506		Blue Bell
2357843		State ex rel. Utility Consumers Council
2414924		City of Boerne at 2170
2414924	Boerne at 2170
2414924		City of Boerne
2410732		Williams
1431414		Memphis Development Foundation
280035	Reece
2925642		Buckner
2330285		Terry
203607		Fustaguio do Nascimento
7124861		Hocker
1433305		Grayson
2411681		Sumner I
901384		Setliff I
901384		Setliff I
1440932		San Giovanni
6596585	Beckwith
6776333		Susser

Time Chart

Generated Files

Branch 1 Output
Branch 2 Output
Full Output CSV

flooie · 2025-04-05T01:09:06Z

Old Data	New Extraction
Inc. v. Toups,	E.g., New Orleans Silversmiths, Inc. v. Toups,
Inc. v. Foy,	Thrift Funds Canal, Inc. v. Foy,
Inc. v. Foy, 261 La. 573,	Thrift Funds Canal, Inc. v. Foy, 261 La. 573,
cit.	cit.
Foy,	Foy,
cit.	cit.
Toups,	Toups,
Odom v. Cherokee Homes, Inc.,	Odom v. Cherokee Homes, Inc.,
Inc. v. I.D.S. Mortgage Corp.,	Slidell Building Supply, Inc. v. I.D.S. Mortgage Corp.,
Inc. v. I.D.S. Mortgage Corp., 273 So.2d 343 (La.App. 1 Cir. 1972), writ den. La.,	La.,
Company v. Subco, Inc.,	of Baton Rouge Bank and Trust Company v. Subco, Inc.,
Foy,	Foy,
cit.	cit.
Union v. Montgomery,	League Central Credit Union v. Montgomery,
Union v. Montgomery, 251 La. 971,	League Central Credit Union v. Montgomery, 251 La. 971,
Monroe v. Carson Homes, Inc.,	Monroe v. Carson Homes, Inc.,
Reed v. Meaux,	Reed v. Meaux,
Corporation v. Breaux,	Chrysler Credit Corporation v. Breaux,
Tapp v. Guaranty Finance Co.,	Tapp v. Guaranty Finance Co.,
Union v. Montgomery, supra, at	Montgomery,
Co. v. Cappel and Curry,	Miller, Lyon and Co. v. Cappel and Curry,
Rocchi v. Keen,	Li Rocchi v. Keen,
Rocchi v. Keen, 242 La. 111,	Li Rocchi v. Keen, 242 La. 111,
Buckner v. Carmack,	Buckner v. Carmack,
Company v. Mones,	Myrtle Grove Packing Company v. Mones,
Company v. Mones, 226 La. 287,	Myrtle Grove Packing Company v. Mones, 226 La. 287,
Co. v. Cappel and Curry,	Co. v. Cappel and Curry,
Rouge v. Heroman,	Louisiana National Bank of Baton Rouge v. Heroman,
Scarborough v. Duke,	Scarborough v. Duke,
Foy,	Foy,
Inc.,	Inc.,
Company v. Soileau,	E.g., Ford Motor Credit Company v. Soileau,
Breaux,	Breaux,
Inc. v. Vicaro,	Polk Chevrolet, Inc. v. Vicaro,
Breaux,	Breaux,
Inc. v. Foy,	Thrift Funds Canal, Inc. v. Foy,
the Court of Appeal (for different reasons),
the Court of Appeal (for different reasons), 261 La. 573,

flooie · 2025-04-05T01:09:41Z

I added a list of old vs new name extraction for a random case. using only plain text.

flooie added 4 commits April 2, 2025 12:54

feat(eyecite): Improve name extraction

f6c6e89

Add check inside method for plaintiff Use that to identify the full plaintiff name Also improve regex building to avoid unnecessary escapes

chore(changes): Update changes.md

441de54

fix(helpers): Use lxml

f839df0

feat(eyecite): Add antecedent guess

6de7ace

Identify antecedent guesses in markup change method name

flooie requested a review from grossir April 2, 2025 17:39

flooie assigned grossir Apr 2, 2025

grossir reviewed Apr 2, 2025

View reviewed changes

eyecite/helpers.py Outdated Show resolved Hide resolved

eyecite/helpers.py Outdated Show resolved Hide resolved

eyecite/helpers.py Outdated Show resolved Hide resolved

refactor(eyecite): Refactor method

2073147

Changed method for speed improvements Added improvements for defendants as well. More example tests Fixed full span

flooie added 4 commits April 3, 2025 13:18

refactor(eyecite): Refactor method

104ddea

Refactor methods for readability Add support for antecedents Strip out stop words if starting extraction

refactor(tests): Add test for pagination

3ed84d8

Pagination has long been in the way and this helps identify cites with pagination in the middle of the citation

chore(tests): Remove unused comments

99d36c2

tests(tests): Add test

58d7dca

Add test to illustrate L.A. support Can now extract defendant out properly

grossir reviewed Apr 3, 2025

View reviewed changes

flooie added 6 commits April 3, 2025 17:11

fixes(eyecite): Fixes

ef814c2

Add Cf. as stop word Dont ovewrite none if we have data Update docstrings and comments

chore(helpers): Docstrings

770b890

tests(tests): Add more tests

d8787f2

chore(tests): Black

49ccc24

tests(tests): Add a few more tests

9d2c329

More edge case examples

chore(eyecite): Lint

fd57cf1

flooie requested a review from grossir April 3, 2025 22:10

grossir reviewed Apr 4, 2025

View reviewed changes

flooie added 2 commits April 4, 2025 10:36

chore(eyecite): Lint

5a9fbd8

Use type ignore and flake8

chore(eyecite): Lint

84e1cdf

flooie added 3 commits April 4, 2025 18:29

fix(tests): Fix failing test

2601a2e

I didnt update a test for our new improved extraction. This one caputres the full text

fix(tests): Add some edge cases

8abc2b5

chore(changes): Update changes.md

f28a5e2

flooie added 2 commits April 4, 2025 20:26

fix(helpers): Check with html and check without it if nothing is found

f1d9ef3

fix(lint): Lint

82f72b0

fix(helpers): Lint

3563f9f

Fix bug in index shift and text cleanup

flooie requested a review from grossir April 5, 2025 01:01

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Markup case names #241

Markup case names #241

flooie commented Apr 2, 2025

grossir left a comment

flooie commented Apr 3, 2025

grossir left a comment •

edited

Loading

grossir Apr 3, 2025

flooie Apr 3, 2025

grossir Apr 4, 2025

grossir Apr 3, 2025

flooie Apr 3, 2025

grossir Apr 3, 2025

flooie Apr 3, 2025

grossir Apr 3, 2025

grossir Apr 3, 2025

flooie commented Apr 3, 2025

grossir left a comment •

edited

Loading

grossir Apr 4, 2025

mlissner commented Apr 4, 2025

flooie commented Apr 4, 2025

flooie commented Apr 5, 2025

flooie commented Apr 5, 2025

github-actions bot commented Apr 5, 2025

flooie commented Apr 5, 2025

flooie commented Apr 5, 2025

Markup case names #241

Are you sure you want to change the base?

Markup case names #241

Conversation

flooie commented Apr 2, 2025

grossir left a comment

Choose a reason for hiding this comment

flooie commented Apr 3, 2025

grossir left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

flooie commented Apr 3, 2025

grossir left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mlissner commented Apr 4, 2025

flooie commented Apr 4, 2025

flooie commented Apr 5, 2025

flooie commented Apr 5, 2025

github-actions bot commented Apr 5, 2025

The Eyecite Report 👁️

Gains and Losses

Time Chart

Generated Files

flooie commented Apr 5, 2025

flooie commented Apr 5, 2025

grossir left a comment •

edited

Loading

grossir left a comment •

edited

Loading